WikiLean · Articles Articles Brain Contribute About

Preliminary: these links have not yet had an independent audit.

Variance

171 passages48 formalized38 partial85 not formalized

Statement only

More ways to filter
Library
About this page

This page

  • Wikipedia revision 1368633150, pinned: the text here is that revision, not today's article.
  • WikiProject Mathematics: B-class, Top importance
  • Annotated article on WikiLean.
  • 171 passages checked, 140 Lean locations (and 78 related declarations).

Central statements: 69 of the 171 passages state the article's subject or a named result or definition; the other 102 are supporting (examples, setup). Of these 69: 21 formalized, 23 partial, 25 not formalized.

What the colours mean

  • Formalized a Lean declaration states it, proved, or a Lean definition defines it
  • Statement only Lean states it, but the proof is not finished
  • Partial Lean states part of it, or a more general or special form
  • Not formalized searched every library and not found

Libraries searched

Searched 48 Lean libraries, pinned 2026-10-17, 3 left out (library set u1).

How these links were made

  • Accuracy, from an audit of a sample of 224 items (blind, not independent): Strong links (exact, close or broad match) on supporting passages that have exactly the right relation: 92% (95% interval 49–99%), from 36 drawn of 1029, weighted by relation and library, an effective sample of 5.1; Links in a counting relation (exact, close, broad, narrow match or part) on supporting passages that have exactly the right relation: 73% (95% interval 40–92%), from 48 drawn of 1740, weighted by relation and library, an effective sample of 8.1; Central and subject passages whose formalization status is right: 86% (95% interval 73–94%), from 44 drawn of 1905, weighted by recorded status, an effective sample of 42; 'Not formalized' on supporting passages that is right (no declaration in the 48 libraries states it): 60% (95% interval 25–87%), from 12 drawn of 1236, weighted by article, an effective sample of 6.1.
  • Audit run tierA-batch1-audit, run by jack, 2 Nov 2026, 21:10 – 3 Nov 2026, 02:10 UTC: judged by Claude Opus 5.5 (claude-opus-5-5).
  • Record set tierA-b1b (cf-locate-recordset/2, statuses by status rule 3), exported 1 Nov 2026 by formal-companion at commit d85385a.
  • Seal sha256 23710337d053e8212378a9e580805628bcdfda61a6fc431e65f24a4a16c9e602.
  • Inventory run inv-tierA-b1r, run by jack, 30 Oct 2026, 07:23–20:00 UTC: worker Claude Opus 5 (claude-opus-5), reviewer Claude Opus 5 (claude-opus-5). It listed the formal passages, each with its kind and prominence, that the locate run then searched.
  • Locate run loc-tierA-b1b, run by jack, 31 Oct 2026, 04:59 – 1 Nov 2026, 03:06 UTC: worker Claude Opus 5 (claude-opus-5), reviewer Claude Opus 5 (claude-opus-5).
  • Challenge run chal-tierA-b1, run by jack, 1 Nov 2026, 03:19–14:25 UTC: worker Claude Opus 5 (claude-opus-5), reviewer Claude Opus 5 (claude-opus-5). In this article it re-checked every Lean location and absence on a central or subject passage (72 locations and 28 absences); on supporting passages, a seeded sample (seed 0; drawn at 0.5 for strong matches, 0.2 for weaker links and 0.25 for absences): 42 of 68 locations and 17 of 61 absences. In all, 114 of 140 Lean locations (81%) and 45 of 89 absences (51%) were re-checked.
  • Passage text: 171 of 171 passages shown in full, each the slice of the pinned revision's text that its record's span sha256 hashes, checked when the store was built.
  • Lean code: each declaration's type, from the declaration corpus of universe u1 (built 18 Oct 2026, sha256 a364fd6aede5), at the commit its card links.

How the text is shown

Each passage is shown in full, as revision 1368633150 states it (checked against its record's span hash when the store was built). Where a passage is only part of a sentence, it is highlighted and the rest of the sentence is shown in grey around it, from the same copy of the revision; a list item also shows the sentence that introduces it. Formulas are drawn from their TeX; hover one to see it.

Each card's Lean code is its declaration's type as Lean prints it, from the declaration corpus of library set u1 at the card's commit, laid out as a header: binders before the colon, then the hypotheses as arrows (the corpus keeps no hypothesis names) and the conclusion, in Mathlib's notation (𝓝, π, R[X], Type*). It is not the source text; Source opens that.

Passages

What “exact match”, “more general” and the other labels mean
  1. Definition · Subject of the article

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    In probability theory and statistics, variance is a measure of dispersion, meaning it is a measure of how far a set of numbers are spread out from their average value.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance more general

    def ProbabilityTheory.variance {Ω : Type*} {mΩ : MeasurableSpace Ω} : (Ω → ℝ) → MeasureTheory.Measure Ω → ℝ

    Mathlib.Probability.Moments.Variance · line 64

    Why more general: Mathlib defines the quantity for an arbitrary measure on Ω through the ENNReal-valued evariance, so it covers both a random variable and a finite collection of numbers under its uniform measure.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 483a2d5c068f · pair key c454dfad0ad4.

    Batteries.RunningStats.variance special case

    def Batteries.RunningStats.variance : Batteries.RunningStats → Float

    Batteries.Data.RunningStats · lines 60–61

    Why a special case: Batteries works with Float values and updates the accumulator by Welford's algorithm.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 4453b52.
    • Location key e361acf0a666 · pair key 9896a7897fd2.
  2. Definition · main statement

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    It is defined as the expected value of the squared deviation from the mean of a random variable.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_eq_integral more general

    theorem ProbabilityTheory.variance_eq_integral {Ω : Type*} {mΩ : MeasurableSpace Ω} {X : Ω → ℝ} {μ : MeasureTheory.Measure Ω} : AEMeasurable X μ → ProbabilityTheory.variance X μ = ∫ (ω : Ω), (X ω - ∫ (x : Ω), X x ∂μ) ^ 2 ∂μ

    Mathlib.Probability.Moments.Variance · lines 157–159

    Why more general: Mathlib states the identity with the mean written as the integral of X and with the measure left as a parameter.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key d4dadf52a412 · pair key 694853583102.

    ProbabilityTheory.variance more general

    def ProbabilityTheory.variance {Ω : Type*} {mΩ : MeasurableSpace Ω} : (Ω → ℝ) → MeasureTheory.Measure Ω → ℝ

    Mathlib.Probability.Moments.Variance · line 64

    Why more general: This is the object the passage defines; Mathlib builds it as the real part of the ENNReal-valued evariance of X.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 3e1970fa99e0 · pair key a4297629e7b8.
  3. Definition · supporting

    Partial

    Partial: Batteries states a special case (a definition).

    The standard deviation is the square root of the variance.

    Show in article · Read it in context on Wikipedia

    Batteries.RunningStats.standardDeviation special case

    def Batteries.RunningStats.standardDeviation : Batteries.RunningStats → Float

    Batteries.Data.RunningStats · lines 70–71

    Why a special case: Batteries builds the square root in Float arithmetic from the accumulator's population variance.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 4453b52.
    • Location key 41b8a41965b9 · pair key f5392b45a426.
  4. Theorem · main statement

    Formalized

    Formalized by 2 parts together (more general):

    1. ProbabilityTheory.centralMoment_two_eq_variance Mathlib (proved)
      • the variance of a random variable is the second central moment of its distribution
    2. ProbabilityTheory.covariance_self Mathlib (proved)
      • the variance of X equals the covariance of X with itself
    Technically, it is the second central moment of a distribution, and the covariance of the random variable with itself, and it is often represented by ⁠ σ2 ⁠, ⁠ s2 ⁠, ⁠ Var(X) ⁠, ⁠ V(X) ⁠, or ⁠ 𝕍(X) ⁠.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.centralMoment_two_eq_variance states part

    theorem ProbabilityTheory.centralMoment_two_eq_variance {Ω : Type*} {m : MeasurableSpace Ω} {X : Ω → ℝ} {μ : MeasureTheory.Measure Ω} : AEMeasurable X μ → ProbabilityTheory.centralMoment X 2 μ = ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Basic · lines 107–108

    Why only part: Mathlib's central moment of order p is the integral of the p-th power of the deviation of X from its mean, and the lemma takes p to be 2.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key ad4d72b2316a · pair key 58593874f416.

    ProbabilityTheory.covariance_self states part

    theorem ProbabilityTheory.covariance_self {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {X : Ω → ℝ} : AEMeasurable X μ → ProbabilityTheory.covariance X X μ = ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 197–201

    Why only part: Mathlib's covariance of two real random variables is the integral of the product of their deviations from their means.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 34ef87ac7573 · pair key 63cf11cd9729.
  5. Notation · supporting

    Formalized

    Formalized: Mathlib states this up to a convention difference (a definition).

    Technically, it is the second central moment of a distribution, and the covariance of the random variable with itself, and it is often represented by ⁠ σ2 ⁠, ⁠ s2 ⁠, ⁠ Var(X) ⁠, ⁠ V(X) ⁠, or ⁠ 𝕍(X) ⁠.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.«termVar[_;_]» close match

    def ProbabilityTheory.«termVar[_;_]» : Lean.ParserDescr

    Mathlib.Probability.Moments.Variance · lines 77–80

    Why close, not exact: Mathlib writes the variance with square brackets and the measure displayed alongside the variable, as Var[X ; measure].

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 468c2641388d · pair key e5f9b3d1c8a2.

    ProbabilityTheory.«termVar[_]» close match

    def ProbabilityTheory.«termVar[_]» : Lean.ParserDescr

    Mathlib.Probability.Moments.Variance · lines 82–85

    Why close, not exact: Mathlib writes Var[X] in square brackets with the measure left to the ambient volume measure.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 7df196ec9aa5 · pair key fca52242e22e.
  6. Theorem · main statement

    Partial

    Partial: Mathlib states a special case (proved).

    An advantage of variance as a measure of dispersion is that it is more amenable to algebraic manipulation than other measures of dispersion such as the expected absolute deviation; for example, the variance of a sum of uncorrelated random variables is equal to the sum of their variances.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.IndepFun.variance_sum special case

    theorem ProbabilityTheory.IndepFun.variance_sum {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {X : ι → Ω → ℝ} {s : Finset ι} : (∀ i ∈ s, MeasureTheory.MemLp (X i) 2 μ) → ((↑s).Pairwise fun i j ↦ ProbabilityTheory.IndepFun (X i) (X j) μ) → ProbabilityTheory.variance (∑ i ∈ s, X i) μ = ∑ i ∈ s, ProbabilityTheory.variance (X i) μ

    Mathlib.Probability.Moments.Variance · lines 424–447

    Why a special case: Mathlib assumes pairwise independence of the summands, a condition that forces the vanishing of their covariances, and the universe holds no declaration about uncorrelated variables as such.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key f5f8ad14365b · pair key e9b023af0356.

    ProbabilityTheory.IndepFun.variance_add special case

    theorem ProbabilityTheory.IndepFun.variance_add {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {X Y : Ω → ℝ} : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.IndepFun X Y μ → ProbabilityTheory.variance (X + Y) μ = ProbabilityTheory.variance X μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 408–415

    Why a special case: This is the two-summand form, again under independence rather than the vanishing of the covariance alone.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 9b5673a4d0c9 · pair key 9d72881dbdee.
  7. Definition · main statement

    Partial

    Partial: Batteries states a special case (a definition).

    If all possible observations of the system are present, then the calculated variance is called the population variance. Normally, however, only a subset is available, and the variance calculated from this is called the sample variance.

    Show in article · Read it in context on Wikipedia

    Batteries.RunningStats.variance special case

    def Batteries.RunningStats.variance : Batteries.RunningStats → Float

    Batteries.Data.RunningStats · lines 60–61

    Why a special case: Batteries names this the population variance in its docstring and computes it in Float from the accumulated count, mean and squared gaps.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 4453b52.
    • Location key efb7180a9a66 · pair key 8bca529516e1.
  8. Definition · main statement

    Partial

    Partial: Batteries states a special case (a definition).

    Normally, however, only a subset is available, and the variance calculated from this is called the sample variance.

    Show in article · Read it in context on Wikipedia

    Batteries.RunningStats.sampleVariance special case

    def Batteries.RunningStats.sampleVariance : Batteries.RunningStats → Float

    Batteries.Data.RunningStats · lines 65–66

    Why a special case: Batteries names this the sample variance and divides the accumulated squared gaps by one less than the count.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 4453b52.
    • Location key 3dfa5632b9c5 · pair key d62c4348609e.

    MetaMathlibExt.initialSegmentSampleVariance special case

    def MetaMathlibExt.initialSegmentSampleVariance : (ℕ → ℕ) → ℕ → ℝ

    MathlibExt.Analysis.Asymptotics.TaylorInitialSampleVariance · line 41

    Why a special case: The declaration fixes the observations to be an integer-valued sequence read along its initial segments and divides by one less than the length.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, low confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 360286bf90ac · pair key 39d15c1454de.
  9. Theorem · supporting

    Not formalized
    If an infinite number of observations are generated using a distribution, then the sample variance calculated from that infinite set will match the value calculated using the distribution's equation for variance.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 63c144a81529).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  10. Construction · supporting

    Not formalized
    A frequency distribution is constructed.The centroid of the distribution gives its mean.A square with sides equal to the difference of each value from the mean is formed for each value.Arranging the squares into a rectangle with one side equal to the number of values, n, results in the other side being the distribution's variance, σ2.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 5 queries and 3 candidates examined (receipt f4f4ac47ccd7).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  11. Definition · main statement · Definition

    Formalized

    Formalized: Mathlib states this exactly (a definition).

    The variance of a random variable X is the expected value of the squared deviation from the mean of ⁠ X ⁠, ⁠ μ = E[X] ⁠: Var(X) = E[(X − μ)2]. This definition encompasses random variables that are generated by processes that are discrete, continuous, neither, or mixed.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.evariance exact match

    def ProbabilityTheory.evariance {Ω : Type*} {mΩ : MeasurableSpace Ω} : (Ω → ℝ) → MeasureTheory.Measure Ω → ENNReal

    Mathlib.Probability.Moments.Variance · line 58

    Note: Values lie in the extended non-negative reals, so a divergent expected squared deviation is recorded as the infinite value, as the passage's expectation would be.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key bae3c7d08259 · pair key d4481e6d80a8.

    ProbabilityTheory.variance close match

    def ProbabilityTheory.variance {Ω : Type*} {mΩ : MeasurableSpace Ω} : (Ω → ℝ) → MeasureTheory.Measure Ω → ℝ

    Mathlib.Probability.Moments.Variance · line 64

    Why close, not exact: Mathlib records the quantity as a real number by taking the real value of the extended-valued version, so a random variable whose expected squared deviation diverges is given the value zero.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 50d3f9ec3e43 · pair key b211dbd35b2f.

    ProbabilityTheory.variance_eq_integral close match

    theorem ProbabilityTheory.variance_eq_integral {Ω : Type*} {mΩ : MeasurableSpace Ω} {X : Ω → ℝ} {μ : MeasureTheory.Measure Ω} : AEMeasurable X μ → ProbabilityTheory.variance X μ = ∫ (ω : Ω), (X ω - ∫ (x : Ω), X x ∂μ) ^ 2 ∂μ

    Mathlib.Probability.Moments.Variance · lines 157–159

    Why close, not exact: States that the library's variance of a random variable is given by the passage's defining integral, with the mean written as the integral of X.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key cb89cf1d4a48 · pair key e47d230fdc9e.
  12. Theorem · supporting · Definition

    Formalized

    Formalized: Mathlib states this exactly (proved).

    The variance can also be thought of as the covariance of a random variable with itself: Var(X) = Cov(X, X).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.covariance_self exact match

    theorem ProbabilityTheory.covariance_self {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {X : Ω → ℝ} : AEMeasurable X μ → ProbabilityTheory.covariance X X μ = ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 197–201

    Note: The equation is stated for a function and a measure without asking the measure to have total mass one, and at the passage's law it is the asserted identity; equality is symmetric, so the order of the two sides carries no difference.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 03ab4cae9f86 · pair key d4c189fdc1a1.
  13. Theorem · supporting · Definition

    Formalized

    With community edits: Formalized. As published it is Not formalized.

    Formalized, by a community edit: Mathlib states a more general result (proved).

    The variance is also equivalent to the second cumulant of a probability distribution that generates ⁠ X ⁠.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_tilted_mul more general

    theorem ProbabilityTheory.variance_tilted_mul {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {X : Ω → ℝ} {t : ℝ} : t ∈ interior (ProbabilityTheory.integrableExpSet X μ) → ProbabilityTheory.variance X (μ.tilted fun x ↦ t * X x) = iteratedDeriv 2 (ProbabilityTheory.cgf X μ) t

    Mathlib.Probability.Moments.Tilted · lines 159–167

    Community Agent The pipeline said related; changed by Meta FAIR 2 (agent) — owned by Jack McCarthy on 9 Nov 2026.

    Why more general: Mathlib has no separate cumulant sequence; the second cumulant is the second derivative at 0 of the cumulant-generating function cgf. This lemma says the variance of X under the measure tilted by t*X equals the second derivative of cgf at t, for any t in the interior of the set where exp(tX) is integrable. At t = 0, tilting a probability measure returns it (tilted_zero), which gives the passage when the mgf is finite near 0.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key ede80ab32450 · pair key 98beb565f2d7.

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 72bb32828b4c).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  14. Notation · supporting · Definition

    Formalized

    Formalized: Mathlib states this up to a convention difference (a definition).

    The variance is typically designated as ⁠ Var(X) ⁠, or sometimes as V(X) or ⁠ 𝕍(X) ⁠, or symbolically as ⁠ σ2X ⁠ or simply σ2 (pronounced "sigma squared").

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.«termVar[_]» close match

    def ProbabilityTheory.«termVar[_]» : Lean.ParserDescr

    Mathlib.Probability.Moments.Variance · lines 82–85

    Why close, not exact: Mathlib writes the variance of X with square brackets and no written measure, taking the ambient measure of the space; the Greek-letter forms the sentence also lists are not introduced.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 8a56a70a2f0e · pair key 73cfe49fd418.

    ProbabilityTheory.«termVar[_;_]» close match

    def ProbabilityTheory.«termVar[_;_]» : Lean.ParserDescr

    Mathlib.Probability.Moments.Variance · lines 77–80

    Why close, not exact: The same bracketed form with the measure written out after a semicolon, so that the distribution the sentence keeps implicit becomes visible.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key f323f4dc738f · pair key 2769258e6355.
  15. Theorem · main statement · Definition

    Formalized

    Formalized: Mathlib states this exactly (proved).

    The expression for the variance can be expanded as follows: Var(X)= E[(X − E[X])2]= E[X2 − 2XE[X] + E[X]2]= E[X2] − 2E[X]E[X] + E[X]2= E[X2] − 2E[X]2 + E[X]2= E[X2] − E[X]2

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_eq_sub exact match

    theorem ProbabilityTheory.variance_eq_sub {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsProbabilityMeasure μ] {X : Ω → ℝ} : MeasureTheory.MemLp X 2 μ → ProbabilityTheory.variance X μ = ∫ (x : Ω), (X ^ 2) x ∂μ - (∫ (x : Ω), X x ∂μ) ^ 2

    Mathlib.Probability.Moments.Variance · lines 226–228

    Note: The integral of the square of X is the mean of the square and the integral of X is the mean, so the equation is the one the passage reaches after expanding; square-integrability is the finiteness the expansion needs in order to split the expectation.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key b9a03a935153 · pair key 8732001f0ece.
  16. Theorem · supporting · Definition

    Formalized

    Formalized: Mathlib states this exactly (proved).

    In other words, the variance of ⁠ X ⁠ is equal to the mean of the square of ⁠ X ⁠ minus the square of the mean of ⁠ X ⁠.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_eq_sub exact match

    theorem ProbabilityTheory.variance_eq_sub {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsProbabilityMeasure μ] {X : Ω → ℝ} : MeasureTheory.MemLp X 2 μ → ProbabilityTheory.variance X μ = ∫ (x : Ω), (X ^ 2) x ∂μ - (∫ (x : Ω), X x ∂μ) ^ 2

    Mathlib.Probability.Moments.Variance · lines 226–228

    Note: Mathlib writes the two means as integrals against X's law, so the equation is the sentence's mean of the square minus the square of the mean.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key c9935e068a3a · pair key 3928123236eb.
  17. Theorem · main statement · Definition › Discrete random variable

    Not formalized
    If the generator of random variable X is discrete with probability mass function x1 ↦ p1, x2 ↦ p2, …, xn ↦ pn, then Var(X) = ∑ni = 1 pi ⋅ (xi − μ)2, where μ is the expected value. That is, μ = ∑ni = 1 pixi. (When such a discrete weighted variance is specified by weights whose sum is not 1, then one divides by the sum of the weights.)

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 0 candidates examined (receipt 1a761f24281b).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  18. Definition · supporting · Definition › Discrete random variable

    Not formalized
    That is, μ = ∑ni = 1 pixi. (When such a discrete weighted variance is specified by weights whose sum is not 1, then one divides by the sum of the weights.)

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 0 candidates examined (receipt 57395dd73160).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  19. Theorem · supporting · Definition › Discrete random variable

    Not formalized
    The variance of a collection of n equally likely values can be written as Var(X) = 1n∑ni = 1(xi − μ)2, where μ is the average value. That is, μ = 1n∑ni = 1 xi.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt 3366e0537b3f).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  20. Theorem · supporting · Definition › Discrete random variable

    Not formalized
    The variance of a set of n equally likely values can be equivalently expressed, without directly referring to the mean, in terms of squared deviations of all pairwise squared distances of points from each other:[2] Var(X) = 1n2∑ni = 1 ∑nj = 1 12(xi − xj)2 = 1n2∑i ∑j > i (xi − xj)2.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 0 candidates examined (receipt 04e16bb06152).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  21. Theorem · main statement · Definition › Absolutely continuous random variable

    Partial

    With community edits: Partial. As published it is Not formalized.

    Partial, by a community edit: Mathlib states part of this (proved).

    If the random variable X has a probability density function ⁠ f(x) ⁠, and F(x) is the corresponding cumulative distribution function, then Var(X) = σ2= ∫ℝ (x − μ)2f(x) dx= ∫ℝ x2f(x) dx − 2μ∫ℝ xf(x) dx + μ2∫ℝ f(x) dx= ∫ℝ x2 dF(x) − 2μ∫ℝ x dF(x) + μ2∫ℝ dF(x)= ∫ℝ x2 dF(x) − 2μ ⋅ μ + μ2 ⋅ 1= ∫ℝ x2 dF(x) − μ2, or equivalently, Var(X) = ∫ℝ x2f(x) dx − μ2, where μ is the expected value of X given by μ = ∫ℝ xf(x) dx = ∫ℝ x dF(x).

    Show in article · Read it in context on Wikipedia

    MeasureTheory.pdf.integral_mul_eq_integral states part

    theorem MeasureTheory.pdf.integral_mul_eq_integral {Ω : Type*} {m : MeasurableSpace Ω} {ℙ : MeasureTheory.Measure Ω} {X : Ω → ℝ} [MeasureTheory.IsFiniteMeasure ℙ] [MeasureTheory.HasPDF X ℙ MeasureTheory.MeasureSpace.volume] : ∫ (x : ℝ), x * (MeasureTheory.pdf X ℙ MeasureTheory.MeasureSpace.volume x).toReal = ∫ (x : Ω), X x ∂ℙ

    Mathlib.Probability.Density · lines 255–258

    Community Agent The pipeline said related; changed by Meta FAIR 2 (agent) — owned by Jack McCarthy on 9 Nov 2026.

    Why only part: States only the passage's final clause: for a real random variable X with a density f with respect to Lebesgue measure, the mean of X equals the integral of x times f(x) over the real line. The variance formulas in terms of f, and the forms written with dF(x), are not stated here. Mathlib proves the clause for any finite measure, which includes the probability case.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 14bb3a986095 · pair key 634c5d931bf1.

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 0 candidates examined (receipt ad855535d960).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  22. Theorem · supporting · Definition › Absolutely continuous random variable

    Not formalized
    If the function x2f(x) is Riemann-integrable on every finite interval [a, b] ⊂ ℝ, then Var(X) = ∫+∞−∞ x2f(x) dx − μ2, where the integral is an improper Riemann integral.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 0 candidates examined (receipt da38307b2fc7).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  23. Definition · supporting · Examples › Exponential distribution

    Formalized

    Formalized: Mathlib states this exactly (a definition).

    The exponential distribution with parameter ⁠ λ > 0 ⁠ is a continuous distribution whose probability density function is given by f(x) = λe−λx on the interval [0, ∞).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.expMeasure exact match

    def ProbabilityTheory.expMeasure : ℝ → MeasureTheory.Measure ℝ

    Mathlib.Probability.Distributions.Exponential · line 94

    Note: At rates that the passage excludes the density is not positive and the measure collapses to zero, a default the passage's mathematics never uses; on the passage's own lambda the two agree.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 0fd0323f0053 · pair key 7cdf76314550.

    TauCeti.Probability.expMeasure_eq_withDensity close match

    theorem TauCeti.Probability.expMeasure_eq_withDensity (r : ℝ) : ProbabilityTheory.expMeasure r = MeasureTheory.MeasureSpace.volume.withDensity (ProbabilityTheory.exponentialPDF r)

    TauCeti.Probability.Distributions.Exponential.Basic · lines 79–80

    Why close, not exact: It records the library's exponential distribution as Lebesgue measure carrying the exponential density, which is the formula by which the passage fixes the distribution.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key fb69965b388a · pair key dfd9aca25205.
  24. Theorem · supporting · Examples › Exponential distribution

    Formalized

    Formalized: TauCeti states this exactly (proved).

    Its mean can be shown to be E[X] = ∫∞0 xλe−λx dx = 1λ.

    Show in article · Read it in context on Wikipedia

    TauCeti.Probability.integral_id_expMeasure exact match

    theorem TauCeti.Probability.integral_id_expMeasure {r : ℝ} : 0 < r → ∫ (x : ℝ), x ∂ProbabilityTheory.expMeasure r = r⁻¹

    TauCeti.Probability.Distributions.Exponential.Basic · lines 138–139

    Note: The distribution is given as Mathlib's exponential measure and the mean as the integral of the identity against it; the law assigns zero mass to the negative half-line, so nothing outside the passage's interval contributes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key cf9601cbfcec · pair key 73754aad9c0b.

    TauCeti.Probability.integral_pow_expMeasure more general

    theorem TauCeti.Probability.integral_pow_expMeasure {r : ℝ} : 0 < r → ∀ (n : ℕ), ∫ (x : ℝ), x ^ n ∂ProbabilityTheory.expMeasure r = ↑n.factorial / r ^ n

    TauCeti.Probability.Distributions.Exponential.Basic · lines 131–134

    Why more general: The declaration gives every moment of the exponential law at once, the passage's mean being the first of them.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 341944bd2d09 · pair key be267f260ec4.
  25. Theorem · supporting · Examples › Exponential distribution

    Formalized

    Formalized: TauCeti states this exactly (proved).

    Using integration by parts and making use of the expected value already calculated, we have: E[X2]= ∫∞0 x2λe−λx dx= [−x2e−λx]∞0 + ∫∞0 2xe−λx dx= 0 + 2λE[X]= 2λ2.

    Show in article · Read it in context on Wikipedia

    TauCeti.Probability.integral_sq_expMeasure exact match

    theorem TauCeti.Probability.integral_sq_expMeasure {r : ℝ} : 0 < r → ∫ (x : ℝ), x ^ 2 ∂ProbabilityTheory.expMeasure r = 2 / r ^ 2

    TauCeti.Probability.Distributions.Exponential.Basic · lines 142–143

    Note: Lean writes the second moment as the integral of the squared identity against the exponential law, which is what the passage's E of X squared denotes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 1f6b0aef046e · pair key 4baab755c2d5.

    TauCeti.Probability.integral_pow_expMeasure more general

    theorem TauCeti.Probability.integral_pow_expMeasure {r : ℝ} : 0 < r → ∀ (n : ℕ), ∫ (x : ℝ), x ^ n ∂ProbabilityTheory.expMeasure r = ↑n.factorial / r ^ n

    TauCeti.Probability.Distributions.Exponential.Basic · lines 131–134

    Why more general: The declaration computes all moments of the exponential law, the passage's second moment among them.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 4c9f188e624b · pair key 5750c2108d2f.
  26. Theorem · main statement · Examples › Exponential distribution

    Formalized

    Formalized: TauCeti states this exactly (proved).

    Thus, the variance of ⁠ X ⁠ is given by Var(X) = E[X2] − E[X]2 = 2λ2 − (1λ)2 = 1λ2.

    Show in article · Read it in context on Wikipedia

    TauCeti.Probability.variance_id_expMeasure exact match

    theorem TauCeti.Probability.variance_id_expMeasure {r : ℝ} : 0 < r → ProbabilityTheory.variance id (ProbabilityTheory.expMeasure r) = (r ^ 2)⁻¹

    TauCeti.Probability.Distributions.Exponential.Basic · lines 147–148

    Note: Lean writes the value as the inverse of the squared rate and reads the variance of the law through the identity function.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key f0f72c1f75b1 · pair key 67b796eb65e5.
  27. Theorem · supporting · Examples › Fair die

    Not formalized
    A fair six-sided die can be modeled as a discrete random variable, ⁠ X ⁠, with outcomes 1 through 6, each with equal probability 1/6. The expected value of ⁠ X ⁠ is (1 + 2 + 3 + 4 + 5 + 6)/6 = 7/2. Therefore, the variance of ⁠ X ⁠ is Var(X)= ∑6i = 1 16(i − 72)2= 16((−5/2)2 + (−3/2)2 + (−1/2)2 + (1/2)2 + (3/2)2 + (5/2)2)= 3512 ≈ 2.92.…

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 45c498a4f547).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  28. Theorem · main statement · Examples › Fair die

    Not formalized
    The expected value of ⁠ X ⁠ is (1 + 2 + 3 + 4 + 5 + 6)/6 = 7/2. Therefore, the variance of ⁠ X ⁠ is Var(X)= ∑6i = 1 16(i − 72)2= 16((−5/2)2 + (−3/2)2 + (−1/2)2 + (1/2)2 + (3/2)2 + (5/2)2)= 3512 ≈ 2.92.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 6dbb799c0dd1).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  29. Theorem · main statement · Examples › Fair die

    Not formalized
    The general formula for the variance of the outcome, ⁠ X ⁠, of an ⁠ n ⁠-sided die is Var(X)= E(X2) − (E(X))2= 1n∑ni = 1 i2 − (1n∑ni = 1 i)2= (n + 1)(2n + 1)6 − (n + 12)2= n2 − 112.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt fa47673323d1).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  30. Theorem · supporting · Examples › Commonly used probability distributions

    Formalized

    Formalized by 2 parts together:

    1. ProbabilityTheory.integral_of_hasLaw_binomial Mathlib (proved)
      • the binomial distribution with parameters n and p has mean n times p
    2. TauCeti.Probability.variance_id_map_cast_binomial TauCeti (proved)
      • the binomial distribution with parameters n and p has variance n times p times (1 - p)
    Binomial distribution Pr (X = k) = (nk)pk(1 − p)n − k np np(1 − p)

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.integral_of_hasLaw_binomial states part

    theorem ProbabilityTheory.integral_of_hasLaw_binomial {Ω : Type*} {m : MeasurableSpace Ω} {P : MeasureTheory.Measure Ω} {n : ℕ} {p : ↑unitInterval} {X : Ω → ℝ} : ProbabilityTheory.HasLaw X (MeasureTheory.Measure.map Nat.cast (ProbabilityTheory.binomial n p)) P → ∫ (x : Ω), X x ∂P = ↑p * ↑n

    Mathlib.Probability.Distributions.Binomial · lines 185–198

    Why only part: Lean records the law as the binomial measure cast from the naturals to the reals, whose mass at k is the passage's binomial coefficient expression, and writes the product in the opposite order.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 6ec43832f922 · pair key d74eedf3f27f.

    TauCeti.Probability.variance_id_map_cast_binomial states part

    theorem TauCeti.Probability.variance_id_map_cast_binomial (n : ℕ) (p : ↑unitInterval) : ProbabilityTheory.variance id (MeasureTheory.Measure.map Nat.cast (ProbabilityTheory.binomial n p)) = ↑p * (1 - ↑p) * ↑n

    TauCeti.Probability.Distributions.Binomial.Basic · lines 180–199

    Why only part: The declaration is stated for the binomial measure pushed forward to the reals, the law whose mass at k is the passage's binomial coefficient expression.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 40b62ec71b77 · pair key 30e301f3b887.
  31. Theorem · supporting · Examples › Commonly used probability distributions

    Partial

    Partial: TauCeti states part of this (proved).

    Geometric distribution Pr (X = k) = (1 − p)k − 1p 1p (1 − p)p2

    Show in article · Read it in context on Wikipedia

    TauCeti.Probability.variance_id_map_cast_geometricMeasure states part

    theorem TauCeti.Probability.variance_id_map_cast_geometricMeasure {p : ↑unitInterval} : ProbabilityTheory.variance id (MeasureTheory.Measure.map Nat.cast (ProbabilityTheory.geometricMeasure p)) = (1 - ↑p) / ↑p ^ 2

    TauCeti.Probability.Distributions.Geometric.Basic · lines 129–137

    Why only part: Lean counts the failures before the first success and so numbers outcomes from zero, while the passage numbers trials from one; the two laws differ by a unit shift, which leaves the spread about the mean untouched.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 591f8cdc6bea · pair key 32637fd7f6b8.
  32. Theorem · supporting · Examples › Commonly used probability distributions

    Formalized

    Formalized by 2 parts together (more general):

    1. ProbabilityTheory.integral_id_gaussianReal Mathlib (proved)
      • the normal distribution with parameters mu and sigma squared has mean mu
    2. ProbabilityTheory.variance_id_gaussianReal Mathlib (proved)
      • the normal distribution with parameters mu and sigma squared has variance sigma squared
    Normal distribution f(x ∣ μ, σ2) = 1√2πσ2e−12(x − μσ)2 μ σ2

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.integral_id_gaussianReal states part

    theorem ProbabilityTheory.integral_id_gaussianReal {μ : ℝ} {v : NNReal} : ∫ (x : ℝ), x ∂ProbabilityTheory.gaussianReal μ v = μ

    Mathlib.Probability.Distributions.Gaussian.Real · lines 535–540

    Why only part: Lean records the mean as the integral of the identity against the law, and lets the second parameter range over all nonnegative reals.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 72b90a3935e8 · pair key a25dc8961579.

    ProbabilityTheory.variance_id_gaussianReal states part

    theorem ProbabilityTheory.variance_id_gaussianReal {μ : ℝ} {v : NNReal} : ProbabilityTheory.variance id (ProbabilityTheory.gaussianReal μ v) = ↑v

    Mathlib.Probability.Distributions.Gaussian.Real · lines 570–571

    Why only part: Lean writes the variance of the identity under the law and keeps the parameter in the nonnegative reals.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 138a68987071 · pair key dc33591ace52.
  33. Theorem · supporting · Examples › Commonly used probability distributions

    Formalized

    Formalized by 2 parts together:

    1. TauCeti.Probability.integral_id_uniformMeasure TauCeti (proved)
      • the continuous uniform distribution on the interval from a to b has mean (a + b)/2
    2. TauCeti.Probability.variance_id_uniformMeasure TauCeti (proved)
      • that uniform distribution has variance the square of (b - a), divided by twelve
    Uniform distribution (continuous) f(x ∣ a, b) = {1b − afor a ≤ x ≤ b,0for x < a or x > b a + b2 (b − a)212

    Show in article · Read it in context on Wikipedia

    TauCeti.Probability.integral_id_uniformMeasure states part

    theorem TauCeti.Probability.integral_id_uniformMeasure {a b : ℝ} : a < b → ∫ (x : ℝ), x ∂TauCeti.Probability.uniformMeasure a b = (a + b) / 2

    TauCeti.Probability.Distributions.Uniform.Basic · lines 284–292

    Why only part: Lean builds the law as normalised Lebesgue measure on the half-open interval, which is the same measure as on the closed one, and states the mean as the integral of the identity.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 3e3c7939910d · pair key 2ed1e39b44b6.

    TauCeti.Probability.variance_id_uniformMeasure states part

    theorem TauCeti.Probability.variance_id_uniformMeasure {a b : ℝ} : a < b → ProbabilityTheory.variance id (TauCeti.Probability.uniformMeasure a b) = (b - a) ^ 2 / 12

    TauCeti.Probability.Distributions.Uniform.Basic · lines 295–306

    Why only part: Lean uses the same normalised Lebesgue measure on the half-open interval and writes the variance of the identity under it.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key fdc4cf4f6fc9 · pair key 50a4d8db92d3.
  34. Theorem · supporting · Examples › Commonly used probability distributions

    Formalized

    Formalized by 2 parts together:

    1. TauCeti.Probability.integral_id_expMeasure TauCeti (proved)
      • the exponential distribution of parameter lambda has mean 1/lambda
    2. TauCeti.Probability.variance_id_expMeasure TauCeti (proved)
      • the exponential distribution of parameter lambda has variance 1 over lambda squared
    Exponential distribution f(x ∣ λ) = λe−λx 1λ 1λ2

    Show in article · Read it in context on Wikipedia

    TauCeti.Probability.integral_id_expMeasure states part

    theorem TauCeti.Probability.integral_id_expMeasure {r : ℝ} : 0 < r → ∫ (x : ℝ), x ∂ProbabilityTheory.expMeasure r = r⁻¹

    TauCeti.Probability.Distributions.Exponential.Basic · lines 138–139

    Why only part: Lean writes the mean as the integral of the identity against the exponential law and spells the value as an inverse rather than a fraction.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key b939f8e5e79a · pair key a35b0a95ec1b.

    TauCeti.Probability.variance_id_expMeasure states part

    theorem TauCeti.Probability.variance_id_expMeasure {r : ℝ} : 0 < r → ProbabilityTheory.variance id (ProbabilityTheory.expMeasure r) = (r ^ 2)⁻¹

    TauCeti.Probability.Distributions.Exponential.Basic · lines 147–148

    Why only part: Lean uses the variance of the identity under the exponential law and spells the value as the inverse of the squared rate.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 785367d0b1dc · pair key 51b4b05e594b.
  35. Theorem · supporting · Examples › Commonly used probability distributions

    Formalized

    Formalized by 2 parts together (more general):

    1. TauCeti.Probability.integral_id_map_cast_poissonMeasure TauCeti (proved)
      • the Poisson distribution of parameter lambda has mean lambda
    2. TauCeti.Probability.variance_id_map_cast_poissonMeasure TauCeti (proved)
      • the Poisson distribution of parameter lambda has variance lambda
    Poisson distribution f(k ∣ λ) = e−λλkk! λ λ

    Show in article · Read it in context on Wikipedia

    TauCeti.Probability.integral_id_map_cast_poissonMeasure states part

    theorem TauCeti.Probability.integral_id_map_cast_poissonMeasure (r : NNReal) : ∫ (x : ℝ), x ∂MeasureTheory.Measure.map Nat.cast (ProbabilityTheory.poissonMeasure r) = ↑r

    TauCeti.Probability.Distributions.Poisson.Basic · lines 85–91

    Why only part: Lean keeps the Poisson law on the naturals and pushes it onto the reals along the cast before integrating the identity, and its rate ranges over all nonnegative reals.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 18ea263b8780 · pair key 932eb4f26502.

    TauCeti.Probability.variance_id_map_cast_poissonMeasure states part

    theorem TauCeti.Probability.variance_id_map_cast_poissonMeasure (r : NNReal) : ProbabilityTheory.variance id (MeasureTheory.Measure.map Nat.cast (ProbabilityTheory.poissonMeasure r)) = ↑r

    TauCeti.Probability.Distributions.Poisson.Basic · lines 127–139

    Why only part: Lean states the variance of the identity under the cast of the Poisson law and lets the rate be any nonnegative real.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key b73d6bcdef28 · pair key 7d3570a8bc39.
  36. Theorem · main statement · Properties › Basic properties

    Formalized

    Formalized: Mathlib states a more general result (proved).

    Variance is non-negative because the squares are positive or zero: Var(X) ≥ 0.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_nonneg more general

    theorem ProbabilityTheory.variance_nonneg {Ω : Type*} {mΩ : MeasurableSpace Ω} (X : Ω → ℝ) (μ : MeasureTheory.Measure Ω) : 0 ≤ ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 204–205

    Why more general: Lean states the inequality for an arbitrary measure on the sample space; a probability law is one such measure and the inequality reads the same in each case.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack, which downgraded it.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 7ff21d0544bd · pair key 39e7b582ff50.
  37. Theorem · supporting · Properties › Basic properties

    Formalized

    With community edits: Formalized. As published it is Partial.

    Formalized, by a community edit: Mathlib states a more general result (proved).

    Variance is non-negative because the squares are positive or zero: Var(X) ≥ 0.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_nonneg more general

    theorem ProbabilityTheory.variance_nonneg {Ω : Type*} {mΩ : MeasurableSpace Ω} (X : Ω → ℝ) (μ : MeasureTheory.Measure Ω) : 0 ≤ ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 204–205

    Community Agent The pipeline said states part; changed by Meta FAIR 2 (agent) — owned by Jack McCarthy on 9 Nov 2026.

    Why more general: The passage's claim is Var(X) >= 0; the remark about squares is its justification. variance_nonneg states that the variance of a real random variable is nonnegative for any measure, not only a probability measure, so the passage is an instance. Mathlib's real-valued variance reads an infinite variance as 0, where the passage would have infinity; the inequality holds either way.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 66ee26b2d3c9 · pair key 117b4bc0fbb5.

    sq_nonneg states part

    theorem sq_nonneg {R : Type u} [Semiring R] [LinearOrder R] [ExistsAddOfLE R] [PosMulMono R] [AddLeftMono R] (a : R) : 0 ≤ a ^ 2

    Mathlib.Algebra.Order.Ring.Unbundled.Basic · lines 608–618

    Why only part: This gives the pointwise sign of the squared deviations but not the conclusion about the variance that the passage goes on to draw.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 7c5c1f07cc36 · pair key 1f74f34ca9cf.
  38. Theorem · main statement · Properties › Basic properties

    Partial

    Partial: Mathlib states a special case (proved).

    The variance of a constant is zero. Var(a) = 0.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_zero special case

    theorem ProbabilityTheory.variance_zero {Ω : Type*} {mΩ : MeasurableSpace Ω} (μ : MeasureTheory.Measure Ω) : ProbabilityTheory.variance 0 μ = 0

    Mathlib.Probability.Moments.Variance · lines 194–195

    Why a special case: Lean records only the constant zero, which is the passage's claim at that one value of a.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 7845ebbf571e · pair key 3be016b3aa67.
  39. Theorem · main statement · Properties › Basic properties

    Formalized

    Formalized: Mathlib states a more general result (proved).

    Conversely, if the variance of a random variable is 0, then it is almost surely a constant. That is, it always has the same value: Var(X) = 0 ⟺ ∃a : P(X = a) = 1.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.evariance_eq_zero_iff more general

    theorem ProbabilityTheory.evariance_eq_zero_iff {Ω : Type*} {mΩ : MeasurableSpace Ω} {X : Ω → ℝ} {μ : MeasureTheory.Measure Ω} : AEMeasurable X μ → (ProbabilityTheory.evariance X μ = 0 ↔ X =ᵐ[μ] fun x ↦ ∫ (x : Ω), X x ∂μ)

    Mathlib.Probability.Moments.Variance · lines 181–184

    Why more general: Lean uses the extended nonnegative-valued variance and writes almost-sure constancy as agreeing almost everywhere with the mean, and it allows any measure rather than only a probability law.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 71f57bc90450 · pair key f5dcc344938f.

    ProbabilityTheory.ae_eq_integral_of_variance_eq_zero states part

    theorem ProbabilityTheory.ae_eq_integral_of_variance_eq_zero {Ω : Type*} {mΩ : MeasurableSpace Ω} {X : Ω → ℝ} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp X 2 μ → ProbabilityTheory.variance X μ = 0 → ∀ᵐ (ω : Ω) ∂μ, X ω = ∫ (x : Ω), X x ∂μ

    Mathlib.Probability.Moments.Variance · lines 162–172

    Why only part: This gives only the direction from vanishing variance to almost-sure constancy, not the converse the passage also asserts.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key fd44a6647553 · pair key b205521c69fb.
  40. Theorem · supporting · Properties › Issues of finiteness

    Formalized

    Formalized: Mathlib states a more general result (proved).

    If a distribution does not have a finite expected value, as is the case for the Cauchy distribution, then the variance cannot be finite either.

    Show in article · Read it in context on Wikipedia

    MeasureTheory.MemLp.integrable more general

    theorem MeasureTheory.MemLp.integrable {α : Type*} {ε : Type*} {m : MeasurableSpace α} {μ : MeasureTheory.Measure α} [TopologicalSpace ε] [ContinuousENorm ε] {q : ENNReal} : 1 ≤ q → ∀ {f : α → ε} [MeasureTheory.IsFiniteMeasure μ], MeasureTheory.MemLp f q μ → MeasureTheory.Integrable f μ

    Mathlib.MeasureTheory.Function.L1Space.Integrable · lines 646–648

    Why more general: Lean renders a finite variance as membership of the square-integrable class and a finite expected value as integrability, and the declaration runs over every exponent above one and any suitably normed codomain, so the article's sentence is its contrapositive at exponent two.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 868945d356b3 · pair key 802242e4300f.
  41. Theorem · supporting · Properties › Issues of finiteness

    Partial

    Partial: TauCeti states part of this (proved).

    However, some distributions may not have a finite variance, despite their expected value being finite.

    Show in article · Read it in context on Wikipedia

    TauCeti.Probability.integrable_id_paretoMeasure_iff states part

    theorem TauCeti.Probability.integrable_id_paretoMeasure_iff {t r : ℝ} : 0 < t → 0 < r → (MeasureTheory.Integrable id (ProbabilityTheory.paretoMeasure t r) ↔ 1 < r)

    TauCeti.Probability.Distributions.Pareto.Basic · lines 166–171

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 7b2acf29024a · pair key 7296eb33625b.

    TauCeti.Probability.not_integrable_sq_paretoMeasure states part

    theorem TauCeti.Probability.not_integrable_sq_paretoMeasure {t r : ℝ} : 0 < t → 0 < r → r ≤ 2 → ¬MeasureTheory.Integrable (fun x ↦ x ^ 2) (ProbabilityTheory.paretoMeasure t r)

    TauCeti.Probability.Distributions.Pareto.Basic · lines 192–195

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 5dbac40d606e · pair key db12901b53e6.
  42. Example · supporting · Properties › Issues of finiteness

    Partial

    Partial: TauCeti states part of this (proved).

    An example is a Pareto distribution whose index k satisfies 1 < k ≤ 2.

    Show in article · Read it in context on Wikipedia

    TauCeti.Probability.integrable_id_paretoMeasure_iff states part

    theorem TauCeti.Probability.integrable_id_paretoMeasure_iff {t r : ℝ} : 0 < t → 0 < r → (MeasureTheory.Integrable id (ProbabilityTheory.paretoMeasure t r) ↔ 1 < r)

    TauCeti.Probability.Distributions.Pareto.Basic · lines 166–171

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key f44f5f510a89 · pair key 2363c710d681.

    TauCeti.Probability.not_integrable_sq_paretoMeasure states part

    theorem TauCeti.Probability.not_integrable_sq_paretoMeasure {t r : ℝ} : 0 < t → 0 < r → r ≤ 2 → ¬MeasureTheory.Integrable (fun x ↦ x ^ 2) (ProbabilityTheory.paretoMeasure t r)

    TauCeti.Probability.Distributions.Pareto.Basic · lines 192–195

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key f806138d88bd · pair key d69ca4eb262d.
  43. Theorem · main statement · Properties › Decomposition

    Formalized

    Formalized: Mathlib states a more general result (proved).

    The general formula for variance decomposition or the law of total variance is: If X and Y are two random variables, and the variance of X exists, then Var[X] = E(Var[X ∣ Y]) + Var(E[X ∣ Y]).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.integral_condVar_add_variance_condExp more general

    theorem ProbabilityTheory.integral_condVar_add_variance_condExp {Ω : Type*} {m₀ m : MeasurableSpace Ω} {X : Ω → ℝ} {μ : MeasureTheory.Measure Ω} : m ≤ m₀ → ∀ [MeasureTheory.IsProbabilityMeasure μ], MeasureTheory.MemLp X 2 μ → ∫ (x : Ω), ProbabilityTheory.condVar m X μ x ∂μ + ProbabilityTheory.variance μ[X | m] μ = ProbabilityTheory.variance X μ

    Mathlib.Probability.CondVar · lines 136–147

    Why more general: The declaration conditions on an arbitrary sub-algebra rather than on a random variable, and the integral against the probability measure is the expectation that the article writes in front of the conditional variance.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 904847056873 · pair key f59b72a9944c.
  44. Definition · supporting · Properties › Decomposition

    Partial

    With community edits: Partial. As published it is Formalized.

    Partial, by a community edit: Mathlib states a special case (a definition).

    Given any particular value y of the random variable Y, there is a conditional expectation E(X ∣ Y = y) given the event Y = y.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.condDistrib special case

    def ProbabilityTheory.condDistrib {α : Type*} {β : Type*} {Ω : Type*} [MeasurableSpace Ω] [StandardBorelSpace Ω] [Nonempty Ω] {x : MeasurableSpace α} [MeasurableSpace β] : (α → Ω) → (α → β) → (μ : MeasureTheory.Measure α) → [MeasureTheory.IsFiniteMeasure μ] → ProbabilityTheory.Kernel β Ω

    Mathlib.Probability.Kernel.CondDistrib · lines 64–66

    Community Agent The pipeline said more general; changed by Meta FAIR 2 (agent) — owned by Jack McCarthy on 9 Nov 2026.

    Why a special case: condDistrib is the regular conditional distribution, a kernel sending each value of the conditioning variable to a probability measure. Evaluated at a set s, it gives the conditional probability of the event given that value, which is the conditional expectation of an indicator: a special case of the passage. It does not define the conditional mean of a general random variable; that is the integral against this kernel.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 8dd8833b63ed · pair key 23acf9b0b375.
  45. Definition · supporting · Properties › Decomposition

    Partial

    With community edits: Partial. As published it is Formalized.

    Partial, by a community edit: Mathlib states a special case (a definition).

    This quantity depends on the particular value y; it is a function g(y) = E(X ∣ Y = y).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.condDistrib special case

    def ProbabilityTheory.condDistrib {α : Type*} {β : Type*} {Ω : Type*} [MeasurableSpace Ω] [StandardBorelSpace Ω] [Nonempty Ω] {x : MeasurableSpace α} [MeasurableSpace β] : (α → Ω) → (α → β) → (μ : MeasureTheory.Measure α) → [MeasureTheory.IsFiniteMeasure μ] → ProbabilityTheory.Kernel β Ω

    Mathlib.Probability.Kernel.CondDistrib · lines 64–66

    Community Agent The pipeline said more general; changed by Meta FAIR 2 (agent) — owned by Jack McCarthy on 9 Nov 2026.

    Why a special case: condDistrib sends each value y of the conditioning variable to the conditional law. At a set s that value is the conditional probability of the event given y, the conditional expectation of an indicator, so it gives the function g only when X is an indicator. For a general X, g(y) is the integral of the identity against this kernel, which condDistrib itself does not define.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 384465f70ab3 · pair key 427bfd46fc2e.
  46. Definition · supporting · Properties › Decomposition

    Partial

    Partial: Mathlib states a more general result, but as a proved theorem, not a definition; a definition passage is formalized only by a Lean definition.

    That same function evaluated at the random variable Y is the conditional expectation ⁠ E(X ∣ Y) = g(Y) ⁠.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.condExp_ae_eq_integral_condDistrib' more general

    theorem ProbabilityTheory.condExp_ae_eq_integral_condDistrib' {α : Type*} {β : Type*} {mα : MeasurableSpace α} {μ : MeasureTheory.Measure α} [MeasureTheory.IsFiniteMeasure μ] {X : α → β} {mβ : MeasurableSpace β} {Ω : Type*} [NormedAddCommGroup Ω] [NormedSpace ℝ Ω] [CompleteSpace Ω] [MeasurableSpace Ω] [BorelSpace Ω] [SecondCountableTopology Ω] {Y : α → Ω} : Measurable X → MeasureTheory.Integrable Y μ → μ[Y | MeasurableSpace.comap X mβ] =ᵐ[μ] fun a ↦ ∫ (y : Ω), y ∂(ProbabilityTheory.condDistrib Y X μ) (X a)

    Mathlib.Probability.Kernel.CondDistrib · lines 382–386

    Why more general: The declaration allows an integrand valued in any complete second countable normed space and states the identity up to a null set, which is how a conditional expectation is pinned down.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key eaecdf1f6d9c · pair key f6bc00f841f9.

    ProbabilityTheory.condExp_ae_eq_integral_condDistrib_id special case

    theorem ProbabilityTheory.condExp_ae_eq_integral_condDistrib_id {β : Type*} {Ω : Type*} {F : Type*} [MeasurableSpace Ω] [StandardBorelSpace Ω] [Nonempty Ω] [NormedAddCommGroup F] {mβ : MeasurableSpace β} [NormedSpace ℝ F] [CompleteSpace F] {X : Ω → β} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsFiniteMeasure μ] : Measurable X → ∀ {f : Ω → F}, MeasureTheory.Integrable f μ → μ[f | MeasurableSpace.comap X mβ] =ᵐ[μ] fun a ↦ ∫ (y : Ω), f y ∂(ProbabilityTheory.condDistrib id X μ) (X a)

    Mathlib.Probability.Kernel.CondDistrib · lines 426–430

    Why a special case: This variant conditions the identity law on the sample space and allows any complete normed codomain for the integrand.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack, which downgraded it.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key aac7f8180018 · pair key 33155b77618c.
  47. Theorem · supporting · Properties › Decomposition

    Not formalized
    In particular, if Y is a discrete random variable assuming possible values y1, y2, y3… with corresponding probabilities p1, p2, p3…,, then in the formula for total variance, the first term on the right-hand side becomes E(Var[X ∣ Y]) = ∑i piσ2i, where σ2i = Var[X ∣ Y = yi].

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 0b58792d3b61).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  48. Theorem · supporting · Properties › Decomposition

    Not formalized
    Similarly, the second term on the right-hand side becomes Var(E[X ∣ Y]) = ∑i piμ2i − (∑i piμi)2 = ∑i piμ2i − μ2, where ⁠ 1 ⁠ and ⁠ μ = ∑i piμi ⁠.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt 56e87cbee4ed).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  49. Theorem · supporting · Properties › Decomposition

    Not formalized
    Thus the total variance is given by Var[X] = ∑i piσ2i + (∑i piμ2i − μ2).

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 86f69f09805c).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  50. Theorem · supporting · Properties › Decomposition

    Not formalized
    A similar formula is applied in analysis of variance, where the corresponding formula is MStotal = MSbetween + MSwithin; here MS refers to the Mean of the Squares.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 3b5effb6ef6f).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  51. Notation · supporting · Properties › Decomposition

    Not formalized
    A similar formula is applied in analysis of variance, where the corresponding formula is MStotal = MSbetween + MSwithin; here MS refers to the Mean of the Squares.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt ead1449f4528).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  52. Theorem · supporting · Properties › Decomposition

    Not formalized
    In linear regression analysis the corresponding formula is MStotal = MSregression + MSresidual.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 1 candidates examined (receipt bc463e80ca47).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  53. Theorem · supporting · Properties › Decomposition

    Not formalized
    This can also be derived from the additivity of variances, since the total (observed) score is the sum of the predicted score and the error score, where the latter two are uncorrelated.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 8e7951e3d983).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  54. Theorem · supporting · Properties › Decomposition

    Not formalized
    Similar decompositions are possible for the sum of squared deviations (sum of squares, SS): SStotal = SSbetween + SSwithin, SStotal = SSregression + SSresidual.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 576548fa6a22).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  55. Theorem · supporting · Properties › Decomposition

    Not formalized
    Similar decompositions are possible for the sum of squared deviations (sum of squares, SS): SStotal = SSbetween + SSwithin, SStotal = SSregression + SSresidual.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 2ccbd1b2d26e).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  56. Theorem · supporting · Properties › Calculation from the CDF

    Not formalized
    The population variance for a non-negative random variable can be expressed in terms of the cumulative distribution function F using 2∫∞0 u(1 − F(u)) du − [∫∞0(1 − F(u)) du]2.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt 3ae6ad52ad2f).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  57. Theorem · main statement · Properties › Characteristic property

    Formalized

    Formalized: TauCeti states a more general result (proved).

    The second moment of a random variable attains the minimum value when taken around the first moment (i.e., mean) of the random variable, i.e. ⁠ argminm E((X − m)2) = E(X) ⁠.

    Show in article · Read it in context on Wikipedia

    TauCeti.setIntegral_norm_sub_setAverage_sq_le_setIntegral_norm_sub_sq more general

    theorem TauCeti.setIntegral_norm_sub_setAverage_sq_le_setIntegral_norm_sub_sq {X : Type*} {F : Type*} [MeasurableSpace X] {μ : MeasureTheory.Measure X} [NormedAddCommGroup F] [InnerProductSpace ℝ F] [CompleteSpace F] {s : Set X} {f : X → F} : μ s ≠ ⊤ → MeasureTheory.IntegrableOn f s μ → ∀ {c : F}, MeasureTheory.IntegrableOn (fun x ↦ ‖f x - c‖ ^ 2) s μ → ∫ (x : X) in s, ‖f x - ⨍ (y : X) in s, f y ∂μ‖ ^ 2 ∂μ ≤ ∫ (x : X) in s, ‖f x - c‖ ^ 2 ∂μ

    TauCeti.MeasureTheory.Integral.Average · lines 274–278

    Why more general: The declaration is stated for an arbitrary real inner product space, an arbitrary set of finite measure and the set average, and the passage is its reading on the real line under a probability law.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key aa812c744a9f · pair key 37a1a161cfab.
  58. Theorem · main statement · Properties › Characteristic property

    Not formalized
    Conversely, if a continuous function φ satisfies argminm E(φ(X − m)) = E(X) for all random variables X, then it is necessarily of the form ⁠ φ(x) = ax2 + b ⁠, where a > 0.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt dec10cbee413).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  59. Theorem · supporting · Properties › Characteristic property

    Partial

    Partial: TauCeti states part of this (proved).

    This also holds in the multidimensional case.

    Show in article · Read it in context on Wikipedia

    TauCeti.setIntegral_norm_sub_setAverage_sq_le_setIntegral_norm_sub_sq states part

    theorem TauCeti.setIntegral_norm_sub_setAverage_sq_le_setIntegral_norm_sub_sq {X : Type*} {F : Type*} [MeasurableSpace X] {μ : MeasureTheory.Measure X} [NormedAddCommGroup F] [InnerProductSpace ℝ F] [CompleteSpace F] {s : Set X} {f : X → F} : μ s ≠ ⊤ → MeasureTheory.IntegrableOn f s μ → ∀ {c : F}, MeasureTheory.IntegrableOn (fun x ↦ ‖f x - c‖ ^ 2) s μ → ∫ (x : X) in s, ‖f x - ⨍ (y : X) in s, f y ∂μ‖ ^ 2 ∂μ ≤ ∫ (x : X) in s, ‖f x - c‖ ^ 2 ∂μ

    TauCeti.MeasureTheory.Integral.Average · lines 274–278

    Why only part: It carries the minimising property over to vector-valued variables but says nothing about which functions of the deviation have that property.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key a7643305ccf2 · pair key 9ba6c486268b.
  60. Theorem · supporting · Properties › Units of measurement

    Not formalized
    In the dice example the standard deviation is √2.9 ≈ 1.7, slightly larger than the expected absolute deviation of 1.5.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 0 candidates examined (receipt 0c8169303e84).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  61. Theorem · main statement · Propagation › Addition and multiplication by a constant

    Formalized

    Formalized: Mathlib states this exactly (proved).

    Variance is invariant with respect to changes in a location parameter. That is, if a constant is added to all values of the variable, the variance is unchanged: Var(X + a) = Var(X).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_add_const exact match

    theorem ProbabilityTheory.variance_add_const {Ω : Type*} {mΩ : MeasurableSpace Ω} {X : Ω → ℝ} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsProbabilityMeasure μ] : MeasureTheory.AEStronglyMeasurable X μ → ∀ (c : ℝ), ProbabilityTheory.variance (fun ω ↦ X ω + c) μ = ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 230–241

    Note: Lean carries the measurability side condition explicitly, which every random variable satisfies.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key f335ff3b53b6 · pair key f656aaa07267.

    ProbabilityTheory.variance_const_add exact match

    theorem ProbabilityTheory.variance_const_add {Ω : Type*} {mΩ : MeasurableSpace Ω} {X : Ω → ℝ} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsProbabilityMeasure μ] : MeasureTheory.AEStronglyMeasurable X μ → ∀ (c : ℝ), ProbabilityTheory.variance (fun ω ↦ c + X ω) μ = ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 243–245

    Note: This is the version with the constant written on the left of the sum.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key f3fb89cd4450 · pair key ad53d07d419a.
  62. Theorem · main statement · Propagation › Addition and multiplication by a constant

    Formalized

    Formalized: Mathlib states a more general result (proved).

    If all values are scaled by a constant, the variance is scaled by the square of that constant: Var(aX) = a2Var(X).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_const_mul more general

    theorem ProbabilityTheory.variance_const_mul {Ω : Type*} {mΩ : MeasurableSpace Ω} (c : ℝ) (X : Ω → ℝ) (μ : MeasureTheory.Measure Ω) : ProbabilityTheory.variance (fun ω ↦ c * X ω) μ = c ^ 2 * ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 207–210

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack, which downgraded it.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 5eac3303a452 · pair key b8d74b8cd78e.

    ProbabilityTheory.variance_mul_const more general

    theorem ProbabilityTheory.variance_mul_const {Ω : Type*} {mΩ : MeasurableSpace Ω} (c : ℝ) (X : Ω → ℝ) (μ : MeasureTheory.Measure Ω) : ProbabilityTheory.variance (fun ω ↦ X ω * c) μ = ProbabilityTheory.variance X μ * c ^ 2

    Mathlib.Probability.Moments.Variance · lines 212–214

    Why more general: This is the version with the scalar written on the right.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack, which downgraded it.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key d94240cd10ab · pair key 2fe1d9db792a.
  63. Theorem · main statement · Propagation › Addition and multiplication by a constant

    Partial

    Partial: Mathlib states part of this (proved).

    The variance of a sum of two random variables is given by Var(aX + bY)= a2Var(X) + b2Var(Y) + 2ab Cov(X, Y)Var(aX − bY)= a2Var(X) + b2Var(Y) − 2ab Cov(X, Y) where Cov(X, Y) is the covariance.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_add states part

    theorem ProbabilityTheory.variance_add {Ω : Type*} {mΩ : MeasurableSpace Ω} {X Y : Ω → ℝ} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.variance (X + Y) μ = ProbabilityTheory.variance X μ + 2 * ProbabilityTheory.covariance X Y μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 263–270

    Why only part: The coefficients a and b of the passage are not carried; the declaration is the case where both equal one.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 72b8c44f55d4 · pair key 50f82d14c665.

    ProbabilityTheory.variance_sub states part

    theorem ProbabilityTheory.variance_sub {Ω : Type*} {mΩ : MeasurableSpace Ω} {X Y : Ω → ℝ} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.variance (X - Y) μ = ProbabilityTheory.variance X μ - 2 * ProbabilityTheory.covariance X Y μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 276–279

    Why only part: As with the sum, the passage's coefficients a and b are specialised to one.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key ce1a2b09302e · pair key 54999551626f.
  64. Theorem · main statement · Propagation › Linear combinations

    Formalized

    Formalized: Mathlib states a more general result (proved).

    In general, for the sum of N random variables ⁠ {X1, …, XN} ⁠, the variance becomes: Var(∑Ni = 1 Xi) = ∑Ni, j = 1 Cov(Xi, Xj) = ∑Ni = 1 Var(Xi) + ∑Ni, j = 1, i ≠ j Cov(Xi, Xj), see also general Bienaymé's identity.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_sum more general

    theorem ProbabilityTheory.variance_sum {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {X : ι → Ω → ℝ} [MeasureTheory.IsFiniteMeasure μ] [Fintype ι] : (∀ (i : ι), MeasureTheory.MemLp (X i) 2 μ) → ProbabilityTheory.variance (∑ i, X i) μ = ∑ i, ∑ j, ProbabilityTheory.covariance (X i) (X j) μ

    Mathlib.Probability.Moments.Variance · lines 295–297

    Why more general: Mathlib writes the right-hand side as one double sum over all index pairs, leaving each equal-index entry as the covariance of a variable with itself rather than listing those entries apart as variances.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack, which downgraded it.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 61be3108ce19 · pair key f7911479fbb8.

    ProbabilityTheory.variance_sum' more general

    theorem ProbabilityTheory.variance_sum' {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {s : Finset ι} {X : ι → Ω → ℝ} [MeasureTheory.IsFiniteMeasure μ] : (∀ i ∈ s, MeasureTheory.MemLp (X i) 2 μ) → ProbabilityTheory.variance (∑ i ∈ s, X i) μ = ∑ i ∈ s, ∑ j ∈ s, ProbabilityTheory.covariance (X i) (X j) μ

    Mathlib.Probability.Moments.Variance · lines 287–293

    Why more general: This is the same identity indexed by an explicit finite set, with the right-hand side kept as a single sum over all pairs of indices.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack, which downgraded it.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key fa0b93ddae51 · pair key 4b9e36bfda0a.
  65. Theorem · main statement · Propagation › Linear combinations

    Partial

    Partial: Mathlib states a special case (proved).

    These results lead to the variance of a linear combination as: Var(∑Ni = 1 aiXi)= ∑Ni, j = 1 aiajCov(Xi, Xj)= ∑Ni = 1 a2iVar(Xi) + ∑i ≠ j aiajCov(Xi, Xj)= ∑Ni = 1 a2iVar(Xi) + 2∑1 ≤ i < j ≤ N aiajCov(Xi, Xj).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_add special case

    theorem ProbabilityTheory.variance_add {Ω : Type*} {mΩ : MeasurableSpace Ω} {X Y : Ω → ℝ} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.variance (X + Y) μ = ProbabilityTheory.variance X μ + 2 * ProbabilityTheory.covariance X Y μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 263–270

    Why a special case: The declaration carries no scalar coefficients, so it realises the passage only at the two-variable unweighted instance.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key f9d4c6fcdc6e · pair key be430a28da3b.

    ProbabilityTheory.variance_smul special case

    theorem ProbabilityTheory.variance_smul {Ω : Type*} {mΩ : MeasurableSpace Ω} (c : ℝ) (X : Ω → ℝ) (μ : MeasureTheory.Measure Ω) : ProbabilityTheory.variance (c • X) μ = c ^ 2 * ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 216–218

    Why a special case: With a single index the cross terms of the passage's formula are absent and only the squared-coefficient term remains.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key cf147f59a49f · pair key 545b918c3ab7.
  66. Definition · main statement · Propagation › Linear combinations

    Not formalized
    If the random variables X1, …, XN are such that Cov(Xi, Xj) = 0 , ∀ (i ≠ j), then they are said to be uncorrelated.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt f3d32b85272e).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  67. Theorem · main statement · Propagation › Linear combinations

    Partial

    Partial: Mathlib states a special case (proved).

    It follows immediately from the expression given earlier that if the random variables X1, …, XN are uncorrelated, then the variance of their sum is equal to the sum of their variances, or, expressed symbolically: Var(∑Ni = 1 Xi) = ∑Ni = 1 Var(Xi).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.IndepFun.variance_sum special case

    theorem ProbabilityTheory.IndepFun.variance_sum {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {X : ι → Ω → ℝ} {s : Finset ι} : (∀ i ∈ s, MeasureTheory.MemLp (X i) 2 μ) → ((↑s).Pairwise fun i j ↦ ProbabilityTheory.IndepFun (X i) (X j) μ) → ProbabilityTheory.variance (∑ i ∈ s, X i) μ = ∑ i ∈ s, ProbabilityTheory.variance (X i) μ

    Mathlib.Probability.Moments.Variance · lines 424–447

    Why a special case: Independence in pairs is assumed rather than the vanishing of the covariances, so only those of the passage's families that are independent are reached.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key e74e1aa88482 · pair key 5e97ccc00259.

    ProbabilityTheory.variance_sum_pi special case

    theorem ProbabilityTheory.variance_sum_pi {ι : Type*} [Fintype ι] {Ω : ι → Type*} {mΩ : (i : ι) → MeasurableSpace (Ω i)} {μ : (i : ι) → MeasureTheory.Measure (Ω i)} [∀ (i : ι), MeasureTheory.IsProbabilityMeasure (μ i)] {X : (i : ι) → Ω i → ℝ} : (∀ (i : ι), MeasureTheory.MemLp (X i) 2 (μ i)) → ProbabilityTheory.variance (∑ i, fun ω ↦ X i (ω i)) (MeasureTheory.Measure.pi μ) = ∑ i, ProbabilityTheory.variance (X i) (μ i)

    Mathlib.Probability.Moments.Variance · lines 449–462

    Why a special case: The family is built as coordinates of a product measure, which is one way the passage's hypothesis can be met.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 3dfb335e5902 · pair key bd46b11f08c3.
  68. Theorem · supporting · Propagation › Linear combinations

    Formalized

    Formalized: Mathlib states this exactly (proved).

    Since independent random variables are always uncorrelated (see Covariance § Uncorrelatedness and independence), the equation above holds in particular when the random variables X1, …, Xn are independent.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.IndepFun.covariance_eq_zero exact match

    theorem ProbabilityTheory.IndepFun.covariance_eq_zero {Ω : Type*} {mΩ : MeasurableSpace Ω} {X Y : Ω → ℝ} {μ : MeasureTheory.Measure Ω} : ProbabilityTheory.IndepFun X Y μ → MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.covariance X Y μ = 0

    Mathlib.Probability.Moments.Covariance · lines 324–332

    Note: Where the second moments are infinite Lean's covariance already returns its default value zero, so the integrability hypotheses leave out only instances in which the conclusion is automatic.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key b7ee6e7e9ef8 · pair key d831d3a5ea5d.
  69. Theorem · main statement · Propagation › Linear combinations

    Formalized

    Formalized: Mathlib states a more general result (proved).

    Since independent random variables are always uncorrelated (see Covariance § Uncorrelatedness and independence), the equation above holds in particular when the random variables X1, …, Xn are independent.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.IndepFun.variance_sum more general

    theorem ProbabilityTheory.IndepFun.variance_sum {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {X : ι → Ω → ℝ} {s : Finset ι} : (∀ i ∈ s, MeasureTheory.MemLp (X i) 2 μ) → ((↑s).Pairwise fun i j ↦ ProbabilityTheory.IndepFun (X i) (X j) μ) → ProbabilityTheory.variance (∑ i ∈ s, X i) μ = ∑ i ∈ s, ProbabilityTheory.variance (X i) μ

    Mathlib.Probability.Moments.Variance · lines 424–447

    Why more general: Mathlib asks that the variables be independent two at a time, so families independent as a whole are covered by it.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key f8b36e7ea7b7 · pair key da1dde212f68.

    ProbabilityTheory.variance_sum_pi special case

    theorem ProbabilityTheory.variance_sum_pi {ι : Type*} [Fintype ι] {Ω : ι → Type*} {mΩ : (i : ι) → MeasurableSpace (Ω i)} {μ : (i : ι) → MeasureTheory.Measure (Ω i)} [∀ (i : ι), MeasureTheory.IsProbabilityMeasure (μ i)] {X : (i : ι) → Ω i → ℝ} : (∀ (i : ι), MeasureTheory.MemLp (X i) 2 (μ i)) → ProbabilityTheory.variance (∑ i, fun ω ↦ X i (ω i)) (MeasureTheory.Measure.pi μ) = ∑ i, ProbabilityTheory.variance (X i) (μ i)

    Mathlib.Probability.Moments.Variance · lines 449–462

    Why a special case: Independence here is built into the product construction rather than assumed of an arbitrary family.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key bc25e2053dd9 · pair key cafb4a898127.
  70. Theorem · supporting · Propagation › Linear combinations

    Partial

    Partial: Mathlib states part of this (proved).

    Thus, independence is sufficient but not necessary for the variance of the sum to equal the sum of the variances.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.IndepFun.variance_sum states part

    theorem ProbabilityTheory.IndepFun.variance_sum {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {X : ι → Ω → ℝ} {s : Finset ι} : (∀ i ∈ s, MeasureTheory.MemLp (X i) 2 μ) → ((↑s).Pairwise fun i j ↦ ProbabilityTheory.IndepFun (X i) (X j) μ) → ProbabilityTheory.variance (∑ i ∈ s, X i) μ = ∑ i ∈ s, ProbabilityTheory.variance (X i) μ

    Mathlib.Probability.Moments.Variance · lines 424–447

    Why only part: It carries the sufficiency half only; nothing in it bears on the passage's claim that independence is not needed.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key fbb2f3004f5f · pair key 475d75eea7d7.
  71. Notation · supporting · Propagation › Linear combinations › Matrix notation for the variance of a linear combination

    Not formalized
    Define X as a column vector of n random variables ⁠ X1, …, Xn ⁠, and c as a column vector of n scalars ⁠ c1, …, cn ⁠.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt 0045ef8bced3).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  72. Definition · supporting · Propagation › Linear combinations › Matrix notation for the variance of a linear combination

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    Therefore, cTX is a linear combination of these random variables, where cT denotes the transpose of ⁠ c ⁠.

    Show in article · Read it in context on Wikipedia

    dotProduct more general

    def dotProduct {m : Type*} {α : Type v} [Fintype m] [Mul α] [AddCommMonoid α] : (m → α) → (m → α) → α

    Mathlib.Data.Matrix.Mul · lines 76–77

    Why more general: This is the defining formula of the object the passage introduces, taken at the reals.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 65063c70220a · pair key 8f303c725857.
  73. Theorem · main statement · Propagation › Linear combinations › Matrix notation for the variance of a linear combination

    Formalized

    Formalized: Mathlib states a more general result (proved).

    Also let Σ be the covariance matrix of ⁠ X ⁠. The variance of cTX is then given by:[4] Var(cTX) = cTΣc.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.covarianceBilin_self more general

    theorem ProbabilityTheory.covarianceBilin_self {E : Type*} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [MeasurableSpace E] [BorelSpace E] {μ : MeasureTheory.Measure E} [CompleteSpace E] [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp id 2 μ → ∀ (x : E), ((ProbabilityTheory.covarianceBilin μ) x) x = ProbabilityTheory.variance (fun u ↦ Inner.inner ℝ x u) μ

    Mathlib.Probability.Moments.CovarianceBilin · lines 78–81

    Why more general: Mathlib carries the covariance of a random vector as a bilinear form on the space instead of as a table of entries, and the two hold the same data.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 44289c628947 · pair key 3f20b50c92d4.

    ProbabilityTheory.covarianceBilinDual_self_eq_variance more general

    theorem ProbabilityTheory.covarianceBilinDual_self_eq_variance {E : Type*} [NormedAddCommGroup E] {mE : MeasurableSpace E} {μ : MeasureTheory.Measure E} [NormedSpace ℝ E] [BorelSpace E] [CompleteSpace E] [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp id 2 μ → ∀ (L : StrongDual ℝ E), ((ProbabilityTheory.covarianceBilinDual μ) L) L = ProbabilityTheory.variance (⇑L) μ

    Mathlib.Probability.Moments.CovarianceBilinDual · lines 355–357

    Why more general: Here the covariance is carried as a form on continuous functionals rather than on vectors, which for the passage's finite-dimensional vector is the same pairing.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 03cf7f7521a4 · pair key 914d303df577.
  74. Theorem · supporting · Propagation › Linear combinations › Matrix notation for the variance of a linear combination

    Formalized

    Formalized: Mathlib states a more general result (proved).

    This implies that the variance of the mean can be written as (with a column vector of ones) Var(x̄) = Var(1n1′X) = 1n21′Σ1.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.covarianceBilin_self more general

    theorem ProbabilityTheory.covarianceBilin_self {E : Type*} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [MeasurableSpace E] [BorelSpace E] {μ : MeasureTheory.Measure E} [CompleteSpace E] [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp id 2 μ → ∀ (x : E), ((ProbabilityTheory.covarianceBilin μ) x) x = ProbabilityTheory.variance (fun u ↦ Inner.inner ℝ x u) μ

    Mathlib.Probability.Moments.CovarianceBilin · lines 78–81

    Why more general: Taking the scalar out of both slots of the covariance form is immediate from its bundled bilinear type, and the form carries the same data as the matrix the passage writes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, low confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 02efb0cff43e · pair key 63855028679d.
  75. Theorem · main statement · Propagation › Sum of variables › Sum of uncorrelated variables

    Partial

    Partial: Mathlib states a special case (proved).

    One reason for the use of the variance in preference to other measures of dispersion is that the variance of the sum (or the difference) of uncorrelated random variables is the sum of their variances: Var(∑ni = 1 Xi) = ∑ni = 1 Var(Xi).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.IndepFun.variance_sum special case

    theorem ProbabilityTheory.IndepFun.variance_sum {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {X : ι → Ω → ℝ} {s : Finset ι} : (∀ i ∈ s, MeasureTheory.MemLp (X i) 2 μ) → ((↑s).Pairwise fun i j ↦ ProbabilityTheory.IndepFun (X i) (X j) μ) → ProbabilityTheory.variance (∑ i ∈ s, X i) μ = ∑ i ∈ s, ProbabilityTheory.variance (X i) μ

    Mathlib.Probability.Moments.Variance · lines 424–447

    Why a special case: Independence is the stronger hypothesis the article itself flags, so this covers the independent families among the passage's uncorrelated ones.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 094c1f0872f8 · pair key f5a0d302c3aa.
  76. Theorem · supporting · Propagation › Sum of variables › Sum of uncorrelated variables

    Partial

    Partial: Mathlib states a special case (proved).

    It is often made with the stronger condition that the variables are independent, but being uncorrelated suffices.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.IndepFun.variance_sum special case

    theorem ProbabilityTheory.IndepFun.variance_sum {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {X : ι → Ω → ℝ} {s : Finset ι} : (∀ i ∈ s, MeasureTheory.MemLp (X i) 2 μ) → ((↑s).Pairwise fun i j ↦ ProbabilityTheory.IndepFun (X i) (X j) μ) → ProbabilityTheory.variance (∑ i ∈ s, X i) μ = ∑ i ∈ s, ProbabilityTheory.variance (X i) μ

    Mathlib.Probability.Moments.Variance · lines 424–447

    Why a special case: No declaration in the universe states the additivity under the weaker vanishing-covariance hypothesis the passage singles out.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 7778e779b4db · pair key b86effa5abf7.
  77. Theorem · main statement · Propagation › Sum of variables › Sum of uncorrelated variables

    Partial

    Partial: TauCeti states a special case (proved).

    So if all the variables have the same variance σ2, then, since division by n is a linear transformation, this formula immediately implies that the variance of their mean is Var(X̄) = Var(1n∑ni = 1 Xi) = 1n2∑ni = 1 Var(Xi) = 1n2nσ2 = σ2n.

    Show in article · Read it in context on Wikipedia

    TauCeti.integral_finiteMean_sub_sq special case

    theorem TauCeti.integral_finiteMean_sub_sq {Ω : Type*} {ι : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} [Fintype ι] [Nonempty ι] [MeasureTheory.IsProbabilityMeasure μ] {Z : ι → Ω → ℝ} {c : ℝ} : (∀ (i : ι), MeasureTheory.MemLp (Z i) 2 μ) → (∀ (i : ι), ∫ (x : Ω), Z i x ∂μ = c) → (Pairwise fun i j ↦ ProbabilityTheory.IndepFun (Z i) (Z j) μ) → ∫ (ω : Ω), ((TauCeti.finiteMean ℝ fun x ↦ Z x ω) - c) ^ 2 ∂μ = (↑(Fintype.card ι) ^ 2)⁻¹ * ∑ i, ∫ (ω : Ω), (Z i ω - c) ^ 2 ∂μ

    TauCeti.Probability.Moments.SampleMean · lines 111–116

    Why a special case: It assumes independence and one shared mean, neither of which the passage requires.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 0cc6bd07615f · pair key d32dedd3845e.

    TauCeti.integral_norm_finiteMean_sub_sq_of_forall_eq special case

    theorem TauCeti.integral_norm_finiteMean_sub_sq_of_forall_eq {Ω : Type*} {ι : Type*} {E : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} […] {X : ι → Ω → E} {m : E} : (∀ (i : ι), MeasureTheory.MemLp (X i) 2 μ) → (∀ (i : ι), ∫ (x : Ω), X i x ∂μ = m) → (Pairwise fun i j ↦ ProbabilityTheory.IndepFun (X i) (X j) μ) → ∀ {γ : ℝ}, (∀ (i : ι), ∫ (ω : Ω), ‖X i ω - m‖ ^ 2 ∂μ = γ) → ∫ (ω : Ω), ‖(TauCeti.finiteMean ℝ fun x ↦ X x ω) - m‖ ^ 2 ∂μ = γ / ↑(Fintype.card ι)
    Full type, as Lean prints it
    ∀ {Ω : Type u_1} {ι : Type u_2} {E : Type u_3} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} [inst : NormedAddCommGroup E] [inst_1 : InnerProductSpace ℝ E] [CompleteSpace E] [inst_3 : MeasurableSpace E] [BorelSpace E] [inst_5 : Fintype ι] [Nonempty ι] [MeasureTheory.IsProbabilityMeasure μ] {X : ι → Ω → E} {m : E}, (∀ (i : ι), MeasureTheory.MemLp (X i) 2 μ) → (∀ (i : ι), ∫ (x : Ω), X i x ∂μ = m) → (Pairwise fun i j ↦ ProbabilityTheory.IndepFun (X i) (X j) μ) → ∀ {γ : ℝ}, (∀ (i : ι), ∫ (ω : Ω), ‖X i ω - m‖ ^ 2 ∂μ = γ) → ∫ (ω : Ω), ‖(TauCeti.finiteMean ℝ fun x ↦ X x ω) - m‖ ^ 2 ∂μ = γ / ↑(Fintype.card ι)

    TauCeti.Probability.Moments.SampleMean · lines 121–129

    Why a special case: It is stated for an arbitrary inner product space and assumes independence and a shared mean on top of the passage's hypotheses.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 44855c0a7205 · pair key b7e4b6cdfcb8.
  78. Theorem · supporting · Propagation › Sum of variables › Sum of uncorrelated variables

    Not formalized
    That is, the variance of the mean decreases when n increases.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 7f680a29d743).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  79. Theorem · supporting · Propagation › Sum of variables › Sum of uncorrelated variables

    Formalized

    Formalized: Mathlib states a more general result (proved).

    To prove the initial statement, it suffices to show that Var(X + Y) = Var(X) + Var(Y).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_add more general

    theorem ProbabilityTheory.variance_add {Ω : Type*} {mΩ : MeasurableSpace Ω} {X Y : Ω → ℝ} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.variance (X + Y) μ = ProbabilityTheory.variance X μ + 2 * ProbabilityTheory.covariance X Y μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 263–270

    Why more general: The declaration keeps the cross term, so it also covers correlated pairs that the passage leaves aside.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 51b2eaff4aef · pair key 9e5f937c5352.

    ProbabilityTheory.IndepFun.variance_add special case

    theorem ProbabilityTheory.IndepFun.variance_add {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {X Y : Ω → ℝ} : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.IndepFun X Y μ → ProbabilityTheory.variance (X + Y) μ = ProbabilityTheory.variance X μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 408–415

    Why a special case: Independence is stronger than the passage's uncorrelatedness, so this covers only part of the passage's pairs.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 8081eb1752ee · pair key ec95249bce0e.
  80. Theorem · supporting · Propagation › Sum of variables › Sum of uncorrelated variables

    Not formalized
    Starting with the definition, Var(X + Y)= E[(X + Y)2] − (E[X + Y])2= E[X2 + 2XY + Y2] − (E[X] + E[Y])2.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 5 queries and 0 candidates examined (receipt f5ade24d70cb).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  81. Theorem · supporting · Propagation › Sum of variables › Sum of uncorrelated variables

    Formalized

    Formalized: Mathlib states a more general result (proved).

    Using the linearity of the expectation operator and the assumption of independence (or uncorrelatedness) of X and Y, this further simplifies as follows: Var(X + Y)= E [X2] + 2E[XY] + E [Y2] − (E[X]2 + 2E[X]E[Y] + E[Y]2)= E[X2] + E[Y2] − E[X]2 − E[Y]2= Var(X) + Var(Y).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_add more general

    theorem ProbabilityTheory.variance_add {Ω : Type*} {mΩ : MeasurableSpace Ω} {X Y : Ω → ℝ} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.variance (X + Y) μ = ProbabilityTheory.variance X μ + 2 * ProbabilityTheory.covariance X Y μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 263–270

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 69ecc65ecb56 · pair key 018b5004a2f2.

    ProbabilityTheory.IndepFun.variance_add special case

    theorem ProbabilityTheory.IndepFun.variance_add {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {X Y : Ω → ℝ} : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.IndepFun X Y μ → ProbabilityTheory.variance (X + Y) μ = ProbabilityTheory.variance X μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 408–415

    Why a special case: Independence is a hypothesis the passage does not require, and independent variables are among the uncorrelated ones the passage speaks of, so this is the passage on those variables (PD4).

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 6882313442d9 · pair key 1e21338a3a7c.
  82. Theorem · main statement · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Partial

    With community edits: Partial. As published it is Not formalized.

    Partial, by a community edit: Mathlib states part of this (statement only: the proof is not finished).

    In general, the variance of the sum of n variables is the sum of their covariances: Var(∑ni = 1 Xi) = ∑ni = 1 ∑nj = 1 Cov(Xi, Xj) = ∑ni = 1 Var(Xi) + 2∑1 ≤ i < j ≤ n Cov(Xi, Xj). (Note: The second equality comes from the fact that Cov(Xi, Xi) = Var(Xi).)

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_sum' states part

    No module or line: an editor named the declaration, not a place in the source.

    Mathlib

    Community Agent Added by Meta FAIR 2 (agent) — owned by Jack McCarthy on 9 Nov 2026

    Why only part: variance_sum' proves that the variance of a finite sum of square-integrable real random variables equals the double sum of their pairwise covariances. That is the passage's first equality. The second form, the sum of the variances plus twice the covariances over i < j, is not stated, so this is partial. Square integrability is the hypothesis that makes these covariances finite.

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 5 queries and 0 candidates examined (receipt d64061e28594).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  83. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Formalized

    Formalized: Mathlib states this exactly (proved).

    …Var(∑ni = 1 Xi) = ∑ni = 1 ∑nj = 1 Cov(Xi, Xj) = ∑ni = 1 Var(Xi) + 2∑1 ≤ i < j ≤ n Cov(Xi, Xj). (Note: The second equality comes from the fact that Cov(Xi, Xi) = Var(Xi).)

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.covariance_self exact match

    theorem ProbabilityTheory.covariance_self {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {X : Ω → ℝ} : AEMeasurable X μ → ProbabilityTheory.covariance X X μ = ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 197–201

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 29b7029527fe · pair key 45f38645c10a.
  84. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Partial

    Partial: Mathlib states a special case (proved).

    Here, Cov(⋅, ⋅) is the covariance, which is zero for independent random variables (if it exists).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.IndepFun.covariance_eq_zero special case

    theorem ProbabilityTheory.IndepFun.covariance_eq_zero {Ω : Type*} {mΩ : MeasurableSpace Ω} {X Y : Ω → ℝ} {μ : MeasureTheory.Measure Ω} : ProbabilityTheory.IndepFun X Y μ → MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.covariance X Y μ = 0

    Mathlib.Probability.Moments.Covariance · lines 324–332

    Why a special case: Mathlib writes the caveat that the covariance exists as square-integrability of each of the two variables, the condition its covariance API always carries.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack, which downgraded it.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 3e8de974238b · pair key 1fa157be8896.
  85. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Formalized

    Formalized: Mathlib states this up to a convention difference (proved).

    The formula states that the variance of a sum is equal to the sum of all elements in the covariance matrix of the components.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_sum close match

    theorem ProbabilityTheory.variance_sum {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {X : ι → Ω → ℝ} [MeasureTheory.IsFiniteMeasure μ] [Fintype ι] : (∀ (i : ι), MeasureTheory.MemLp (X i) 2 μ) → ProbabilityTheory.variance (∑ i, X i) μ = ∑ i, ∑ j, ProbabilityTheory.covariance (X i) (X j) μ

    Mathlib.Probability.Moments.Variance · lines 295–297

    Why close, not exact: Mathlib keeps the covariance matrix as the indexed family of covariances and totals it with an iterated sum over both indices.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 7030559a3d41 · pair key 730dd290d0a7.

    ProbabilityTheory.variance_sum' close match

    theorem ProbabilityTheory.variance_sum' {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {s : Finset ι} {X : ι → Ω → ℝ} [MeasureTheory.IsFiniteMeasure μ] : (∀ i ∈ s, MeasureTheory.MemLp (X i) 2 μ) → ProbabilityTheory.variance (∑ i ∈ s, X i) μ = ∑ i ∈ s, ∑ j ∈ s, ProbabilityTheory.covariance (X i) (X j) μ

    Mathlib.Probability.Moments.Variance · lines 287–293

    Why close, not exact: Mathlib indexes the summands by a finite set and totals the covariances of all ordered pairs drawn from it.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 5c34cce6447c · pair key 7e4f8ff5731d.
  86. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Not formalized
    The next expression states equivalently that the variance of the sum is the sum of the diagonal of covariance matrix plus two times the sum of its upper triangular elements (or its lower triangular elements); this emphasizes that the covariance matrix is symmetric.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 0 candidates examined (receipt 20daf998d8bc).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  87. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Formalized

    Formalized: TauCeti states a more general result (proved).

    The next expression states equivalently that the variance of the sum is the sum of the diagonal of covariance matrix plus two times the sum of its upper triangular elements (or its lower triangular elements); this emphasizes that the covariance matrix is symmetric.

    Show in article · Read it in context on Wikipedia

    TauCeti.covMatrix_apply_comm more general

    theorem TauCeti.covMatrix_apply_comm {ι : Type*} (μ : MeasureTheory.Measure (EuclideanSpace ℝ ι)) (i j : ι) : TauCeti.covMatrix μ i j = TauCeti.covMatrix μ j i

    TauCeti.Probability.Moments.Covariance · lines 61–63

    Why more general: The declaration is stated for an arbitrary index type and for measures whose second moments need not exist, so it covers more than the passage's finite covariance matrix.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key de6c00b9c571 · pair key 976515f2fbb0.

    TauCeti.isHermitian_covMatrix more general

    theorem TauCeti.isHermitian_covMatrix {ι : Type*} (μ : MeasureTheory.Measure (EuclideanSpace ℝ ι)) : (TauCeti.covMatrix μ).IsHermitian

    TauCeti.Probability.Moments.Covariance · lines 66–69

    Why more general: The declaration is stated for an arbitrary index type and every measure on the Euclidean space, and phrases the symmetry as a Hermitian condition.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key c99b321b3fd0 · pair key fc89d8a33d7d.
  88. Theorem · main statement · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Partial

    Partial: TauCeti states a special case (proved).

    So, if the variables have equal variance σ2 and the average correlation of distinct variables is ρ, then the variance of their mean is Var(X̄) = σ2n + n − 1nρσ2.

    Show in article · Read it in context on Wikipedia

    TauCeti.Probability.Contractable.variance_blockAverage special case

    theorem TauCeti.Probability.Contractable.variance_blockAverage {Ω : Type*} [MeasurableSpace Ω] {μ : MeasureTheory.Measure Ω} {X : ℕ → Ω → ℝ} [MeasureTheory.IsFiniteMeasure μ] : TauCeti.Probability.Contractable μ X → (∀ (n : ℕ), MeasureTheory.MemLp (X n) 2 μ) → ∀ {n : ℕ}, 0 < n → ∀ {k : Fin n → ℕ}, Function.Injective k → ProbabilityTheory.variance (TauCeti.Probability.blockAverage X k) μ = (ProbabilityTheory.variance (X 0) μ - ProbabilityTheory.covariance (X 0) (X 1) μ) / ↑n + ProbabilityTheory.covariance (X 0) (X 1) μ

    TauCeti.Probability.Exchangeability.L2.BlockAverages · lines 98–107

    Why a special case: Its contractability hypothesis makes all pair covariances coincide, so the passage restricted to exchangeable sequences is what it states; the passage itself allows unequal correlations whose average is rho.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 43b997456991 · pair key 055e8920fd24.
  89. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Not formalized
    This implies that the variance of the mean increases with the average of the correlations.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 473f9fb89d7a).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  90. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Partial

    Partial: TauCeti states a special case (proved).

    Moreover, if the variables have unit variance, for example if they are standardized, then this simplifies to Var(X̄) = 1n + n − 1nρ.

    Show in article · Read it in context on Wikipedia

    TauCeti.Probability.Contractable.variance_blockAverage special case

    theorem TauCeti.Probability.Contractable.variance_blockAverage {Ω : Type*} [MeasurableSpace Ω] {μ : MeasureTheory.Measure Ω} {X : ℕ → Ω → ℝ} [MeasureTheory.IsFiniteMeasure μ] : TauCeti.Probability.Contractable μ X → (∀ (n : ℕ), MeasureTheory.MemLp (X n) 2 μ) → ∀ {n : ℕ}, 0 < n → ∀ {k : Fin n → ℕ}, Function.Injective k → ProbabilityTheory.variance (TauCeti.Probability.blockAverage X k) μ = (ProbabilityTheory.variance (X 0) μ - ProbabilityTheory.covariance (X 0) (X 1) μ) / ↑n + ProbabilityTheory.covariance (X 0) (X 1) μ

    TauCeti.Probability.Exchangeability.L2.BlockAverages · lines 98–107

    Why a special case: For standardized coordinates the common covariance is exactly the correlation rho, and the declaration's exchangeability hypothesis is more than the passage asks, which only fixes the average correlation.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 6fe5f01f1543 · pair key 1209315c7785.
  91. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Not formalized
    This converges to ρ if n goes to infinity, provided that the average correlation remains constant or converges too.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt b79182c16f42).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  92. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Not formalized
    So for the variance of the mean of standardized variables with equal correlations or converging average correlation we have limn → ∞ Var(X̄) = ρ.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 57d770486631).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  93. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Not formalized
    Therefore, the variance of the mean of a large number of standardized variables is approximately equal to their average correlation.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 1 candidates examined (receipt 792549eaf53b).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  94. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of correlated variables with fixed sample size

    Partial

    Partial: Mathlib states part of this (proved).

    This makes clear that the sample mean of correlated variables does not generally converge to the population mean, even though the law of large numbers states that the sample mean will converge for independent variables.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.strong_law_ae_real states part

    theorem ProbabilityTheory.strong_law_ae_real {Ω : Type*} {m : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} (X : ℕ → Ω → ℝ) : MeasureTheory.Integrable (X 0) μ → Pairwise (Function.onFun (fun x1 x2 ↦ ProbabilityTheory.IndepFun x1 x2 μ) X) → (∀ (i : ℕ), ProbabilityTheory.IdentDistrib (X i) (X 0) μ μ) → ∀ᵐ (ω : Ω) ∂μ, Filter.Tendsto (fun n ↦ (∑ i ∈ Finset.range n, X i ω) / ↑n) Filter.atTop (𝓝 (∫ (x : Ω), X 0 x ∂μ))

    Mathlib.Probability.StrongLaw · lines 598–631

    Why only part: It carries the independent half of the passage; the failure of convergence for correlated observations is not stated anywhere in the universe.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 112f31957300 · pair key 62e161629009.

    ProbabilityTheory.strong_law_ae states part

    theorem ProbabilityTheory.strong_law_ae {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [CompleteSpace E] [MeasurableSpace E] [BorelSpace E] (X : ℕ → Ω → E) : MeasureTheory.Integrable (X 0) μ → Pairwise (Function.onFun (fun x1 x2 ↦ ProbabilityTheory.IndepFun x1 x2 μ) X) → (∀ (i : ℕ), ProbabilityTheory.IdentDistrib (X i) (X 0) μ μ) → ∀ᵐ (ω : Ω) ∂μ, Filter.Tendsto (fun n ↦ (↑n)⁻¹ • ∑ i ∈ Finset.range n, X i ω) Filter.atTop (𝓝 (∫ (x : Ω), X 0 x ∂μ))

    Mathlib.Probability.StrongLaw · lines 787–818

    Why only part: This is the version for values in a general Banach space, so it stands behind the real-valued statement the passage is about.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 3649db55faa6 · pair key 221f3aa6d5cf.
  95. Theorem · main statement · Propagation › Sum of variables › Sum of correlated variables › Sum of uncorrelated variables with random sample size

    Not formalized
    In such cases, the sample size N is a random variable whose variation adds to the variation of X, such that,[8] Var(∑Ni = 1 Xi) = E[N]Var(X) + Var(N)(E[X])2 which follows from the law of total variance.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 5 queries and 3 candidates examined (receipt a64dda290b19).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  96. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of uncorrelated variables with random sample size

    Not formalized
    If N has a Poisson distribution, then E[N] = Var(N) with estimator n = N.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 0 candidates examined (receipt 0bef27738e9d).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  97. Theorem · supporting · Propagation › Sum of variables › Sum of correlated variables › Sum of uncorrelated variables with random sample size

    Not formalized
    If N has a Poisson distribution, then E[N] = Var(N) with estimator n = N. So, the estimator of Var(∑ni = 1 Xi) becomes ⁠ nSx2 + nX̄2 ⁠, giving SE(X̄) = √(Sx2 + X̄2)/n (see Standard error § Standard error of the sample mean).

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 0 candidates examined (receipt 1a12841e74f4).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  98. Theorem · main statement · Propagation › Sum of variables › Weighted sum of variables

    Partial

    Partial: Mathlib states a special case (proved).

    The scaling property and the Bienaymé formula, along with the property of the covariance Cov(aX, bY) = ab Cov(X, Y) jointly imply that Var(aX ± bY) = a2Var(X) + b2Var(Y) ± 2ab Cov(X, Y).

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_add special case

    theorem ProbabilityTheory.variance_add {Ω : Type*} {mΩ : MeasurableSpace Ω} {X Y : Ω → ℝ} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.variance (X + Y) μ = ProbabilityTheory.variance X μ + 2 * ProbabilityTheory.covariance X Y μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 263–270

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 509e5e339a7b · pair key 8b91b70a8f5b.

    ProbabilityTheory.variance_sub special case

    theorem ProbabilityTheory.variance_sub {Ω : Type*} {mΩ : MeasurableSpace Ω} {X Y : Ω → ℝ} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.variance (X - Y) μ = ProbabilityTheory.variance X μ - 2 * ProbabilityTheory.covariance X Y μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 276–279

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key c9cc1d585ba1 · pair key 238d47a87110.

    ProbabilityTheory.variance_smul special case

    theorem ProbabilityTheory.variance_smul {Ω : Type*} {mΩ : MeasurableSpace Ω} (c : ℝ) (X : Ω → ℝ) (μ : MeasureTheory.Measure Ω) : ProbabilityTheory.variance (c • X) μ = c ^ 2 * ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 216–218

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key dfa449912211 · pair key 4f6c00812493.
  99. Example · supporting · Propagation › Sum of variables › Weighted sum of variables

    Not formalized
    For example, if X and Y are uncorrelated and the weight of X is two times the weight of Y, then the weight of the variance of X will be four times the weight of the variance of Y.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 0 candidates examined (receipt d4536c0f078b).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  100. Theorem · main statement · Propagation › Sum of variables › Weighted sum of variables

    Partial

    Partial: Mathlib states a special case (proved).

    The expression above can be extended to a weighted sum of multiple variables: Var(∑ni aiXi) = ∑ni = 1 a2iVar(Xi) + 2∑1 ≤ i ∑< j ≤ n aiajCov(Xi, Xj)

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_sum' special case

    theorem ProbabilityTheory.variance_sum' {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {s : Finset ι} {X : ι → Ω → ℝ} [MeasureTheory.IsFiniteMeasure μ] : (∀ i ∈ s, MeasureTheory.MemLp (X i) 2 μ) → ProbabilityTheory.variance (∑ i ∈ s, X i) μ = ∑ i ∈ s, ∑ j ∈ s, ProbabilityTheory.covariance (X i) (X j) μ

    Mathlib.Probability.Moments.Variance · lines 287–293

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key d8d2b0eb2bbf · pair key f1206e6fcc22.

    ProbabilityTheory.variance_sum special case

    theorem ProbabilityTheory.variance_sum {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {X : ι → Ω → ℝ} [MeasureTheory.IsFiniteMeasure μ] [Fintype ι] : (∀ (i : ι), MeasureTheory.MemLp (X i) 2 μ) → ProbabilityTheory.variance (∑ i, X i) μ = ∑ i, ∑ j, ProbabilityTheory.covariance (X i) (X j) μ

    Mathlib.Probability.Moments.Variance · lines 295–297

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key d5d48ac6fe63 · pair key 46e26ead9342.

    ProbabilityTheory.IndepFun.variance_sum special case

    theorem ProbabilityTheory.IndepFun.variance_sum {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {ι : Type*} {X : ι → Ω → ℝ} {s : Finset ι} : (∀ i ∈ s, MeasureTheory.MemLp (X i) 2 μ) → ((↑s).Pairwise fun i j ↦ ProbabilityTheory.IndepFun (X i) (X j) μ) → ProbabilityTheory.variance (∑ i ∈ s, X i) μ = ∑ i ∈ s, ProbabilityTheory.variance (X i) μ

    Mathlib.Probability.Moments.Variance · lines 424–447

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 6b15eaf83cef · pair key 87f7abfc6bed.
  101. Theorem · main statement · Propagation › Product of variables › Product of independent variables

    Not formalized
    If two variables X and Y are independent, the variance of their product is given by[9] Var(XY) = [E(X)]2Var(Y) + [E(Y)]2Var(X) + Var(X)Var(Y).

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt 132f470193b7).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  102. Theorem · supporting · Propagation › Product of variables › Product of independent variables

    Not formalized
    Equivalently, using the basic properties of expectation, it is given by Var(XY) = E(X2)E(Y2) − [E(X)]2[E(Y)]2.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt fe82f948113f).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  103. Theorem · main statement · Propagation › Product of variables › Product of statistically dependent variables

    Partial

    With community edits: Partial. As published it is Not formalized.

    Partial, by a community edit: Mathlib states part of this (proved).

    In general, if two variables are statistically dependent, then the variance of their product is given by: Var(XY) =E[X2Y2] − [E(XY)]2=Cov(X2, Y2) + E(X2)E(Y2) − [E(XY)]2=Cov(X2, Y2) + (Var(X) + [E(X)]2)(Var(Y) + [E(Y)]2)− [Cov(X, Y) + E(X)E(Y)]2

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_eq_sub states part

    theorem ProbabilityTheory.variance_eq_sub {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsProbabilityMeasure μ] {X : Ω → ℝ} : MeasureTheory.MemLp X 2 μ → ProbabilityTheory.variance X μ = ∫ (x : Ω), (X ^ 2) x ∂μ - (∫ (x : Ω), X x ∂μ) ^ 2

    Mathlib.Probability.Moments.Variance · lines 226–228

    Community Agent The pipeline said related; changed by Meta FAIR 2 (agent) — owned by Jack McCarthy on 9 Nov 2026.

    Why only part: variance_eq_sub says Var[Z] = E[Z^2] - E[Z]^2 for any square-integrable Z on a probability space. With Z = XY this is the passage's first equality, Var(XY) = E[X^2 Y^2] - E(XY)^2 (since (XY)^2 = X^2 Y^2 pointwise). The later rewrites in terms of Cov(X^2,Y^2) and Cov(X,Y) follow from covariance_eq_sub and variance_eq_sub and are not stated as one lemma about products.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key d3b266e3e8a7 · pair key 439d48dfb365.

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 0 candidates examined (receipt 9eaf0862b57c).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  104. Theorem · main statement · Propagation › Arbitrary functions

    Partial

    Partial: Mathlib states a special case (proved).

    For example, the approximate variance of a function of one variable is given by Var[f(X)] ≈ (f′(E[X]))2Var[X] provided that f is twice differentiable and that the mean and variance of X are finite.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_const_mul special case

    theorem ProbabilityTheory.variance_const_mul {Ω : Type*} {mΩ : MeasurableSpace Ω} (c : ℝ) (X : Ω → ℝ) (μ : MeasureTheory.Measure Ω) : ProbabilityTheory.variance (fun ω ↦ c * X ω) μ = c ^ 2 * ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 207–210

    Why a special case: The declaration covers maps that scale by a constant, where the derivative at the mean is that constant and the article's approximation becomes an equality; it says nothing about a curved function or about the size of the error the article leaves implicit.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, low confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key a96337ee958e · pair key 3189ea397edc.
  105. Definition · supporting · Population variance and sample variance

    Not formalized
    The estimator is a function of the sample of n observations drawn without observational bias from the whole population of potential observations. In this example, the sample would be the set of actual measurements of yesterday's rainfall from available rain gauges within the geography of interest.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt d88f8a24a54f).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  106. Definition · main statement · Population variance and sample variance

    Partial

    Partial: Mathlib states part of this (proved).

    The simplest estimators for population mean and population variance are simply the mean and variance of the sample, the sample mean and (uncorrected) sample variance – these are consistent estimators (they converge to the value of the whole population as the number of samples increases) but can be improved.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.strong_law_ae_real states part

    theorem ProbabilityTheory.strong_law_ae_real {Ω : Type*} {m : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} (X : ℕ → Ω → ℝ) : MeasureTheory.Integrable (X 0) μ → Pairwise (Function.onFun (fun x1 x2 ↦ ProbabilityTheory.IndepFun x1 x2 μ) X) → (∀ (i : ℕ), ProbabilityTheory.IdentDistrib (X i) (X 0) μ μ) → ∀ᵐ (ω : Ω) ∂μ, Filter.Tendsto (fun n ↦ (∑ i ∈ Finset.range n, X i ω) / ↑n) Filter.atTop (𝓝 (∫ (x : Ω), X 0 x ∂μ))

    Mathlib.Probability.StrongLaw · lines 598–631

    Why only part: The declaration speaks of the average of the observations only; the behaviour of the sample variance as the sample grows is not among the things it says.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key a6e814cf5929 · pair key 4b3d13cb27dc.
  107. Definition · main statement · Population variance and sample variance

    Partial

    With community edits: Partial. As published it is Formalized.

    Partial, by a community edit: Batteries states something related (a definition).

    Most simply, the sample variance is computed as the sum of squared deviations about the (sample) mean, divided by n as the number of samples. However, using values other than n improves the estimator in various ways.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance states part

    No module or line: an editor named the declaration, not a place in the source.

    Mathlib

    Community Agent Added by Meta FAIR 2 (agent) — owned by Jack McCarthy on 9 Nov 2026

    Why only part: ProbabilityTheory.variance is the variance of a real random variable X under a measure: the integral of (X - E[X])^2, as a real number. Give the n sample indices the uniform probability measure and let X send index i to the i-th observation: E[X] is the sample mean and the variance is the sum of squared deviations about it divided by n. The passage's remark that other divisors improve the estimator is not covered.

  108. Theorem · main statement · Population variance and sample variance

    Partial

    Partial: Batteries states part of this (a definition).

    Four common values for the denominator are n, n − 1, n + 1, and n − 1.5: n is the simplest (the variance of the sample), n − 1 eliminates bias,[10] n + 1 minimizes mean squared error for the normal distribution,[11] and n − 1.5 mostly eliminates bias in unbiased estimation of standard deviation for the normal distribution.

    Show in article · Read it in context on Wikipedia

    Batteries.RunningStats.variance states part

    def Batteries.RunningStats.variance : Batteries.RunningStats → Float

    Batteries.Data.RunningStats · lines 60–61

    Why only part: Lean stores the sample in a Welford accumulator of floating point numbers and returns its squared-deviation field divided by the count, which the library documents as the variance of the data itself.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 4453b52.
    • Location key 275b8bedde63 · pair key 301e1cd12ece.
  109. Theorem · main statement · Population variance and sample variance

    Not formalized
    Firstly, if the true population mean is unknown, then the sample variance (which uses the sample mean in place of the true mean) is a biased estimator: it underestimates the variance by a factor of (n − 1) / n; correcting this factor, resulting in the sum of squared deviations about the sample mean divided by n − 1 instead of n, is called Bessel's correction.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt 7a5bf8f834c6).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  110. Definition · main statement · Population variance and sample variance

    Partial

    With community edits: Partial. As published it is Formalized.

    Partial, by a community edit: Batteries states something related (a definition).

    Firstly, if the true population mean is unknown, then the sample variance (which uses the sample mean in place of the true mean) is a biased estimator: it underestimates the variance by a factor of (n − 1) / n; correcting this factor, resulting in the sum of squared deviations about the sample mean divided by n − 1 instead of n, is called Bessel's correction.[10]

    Show in article · Read it in context on Wikipedia

    MetaMathlibExt.initialSegmentSampleVariance special case

    def MetaMathlibExt.initialSegmentSampleVariance : (ℕ → ℕ) → ℕ → ℝ

    MathlibExt.Analysis.Asymptotics.TaylorInitialSampleVariance · line 41

    Why a special case: The declaration fixes observations valued in the natural numbers and indexed from one, where the article's data are arbitrary measurements, and it inlines the sample mean into the formula.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key a2712c518bcf · pair key 9e9cb90ba349.
  111. Definition · main statement · Population variance and sample variance

    Partial

    With community edits: Partial. As published it is Not formalized.

    Partial, by a community edit: ATLAS (atlas-lean) states a special case (a definition).

    The resulting estimator is unbiased and is called the (corrected) sample variance or unbiased sample variance.

    Show in article · Read it in context on Wikipedia

    MetaMathlibExt.initialSegmentSampleVariance special case

    def MetaMathlibExt.initialSegmentSampleVariance : (ℕ → ℕ) → ℕ → ℝ

    MathlibExt.Analysis.Asymptotics.TaylorInitialSampleVariance · line 41

    Community Agent The pipeline said related; changed by Meta FAIR 2 (agent) — owned by Jack McCarthy on 9 Nov 2026.

    Why a special case: initialSegmentSampleVariance defines the corrected sample variance (1/(n-1)) times the sum of (a_j - m)^2 over j = 1..n, where m is the sample mean of the first n terms. Its docstring calls it the unbiased sample variance. It only covers data given as a natural-number sequence, so it is a special case of the passage's estimator for general real samples. It does not state unbiasedness.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack, which downgraded it.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 34343f3dbccb · pair key 27f41c602b0b.

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt ae6d607e246c).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  112. Theorem · supporting · Population variance and sample variance

    Not formalized
    If the mean is determined in some other way than from the same samples used to estimate the variance, then this bias does not arise, and the variance can safely be estimated as that of the samples about the (independently known) mean.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt c717dbaa6682).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  113. Theorem · main statement · Population variance and sample variance

    Not formalized
    Secondly, the sample variance does not generally minimize mean squared error between sample variance and population variance.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 79e479058770).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  114. Theorem · supporting · Population variance and sample variance

    Not formalized
    Correcting for bias often makes this worse: one can always choose a scale factor that performs better than the corrected sample variance, though the optimal scale factor depends on the excess kurtosis of the population (see Mean squared error § Variance) and introduces bias.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 1 candidates examined (receipt 6fa864ef3b76).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  115. Theorem · supporting · Population variance and sample variance

    Not formalized
    This always consists of scaling down the unbiased estimator (dividing by a number larger than n − 1) and is a simple example of a shrinkage estimator: one "shrinks" the unbiased estimator towards zero.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 1 candidates examined (receipt d0e08f012dbf).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  116. Theorem · main statement · Population variance and sample variance

    Not formalized
    For the normal distribution, dividing by n + 1 (instead of n − 1 or n) minimizes mean squared error.[11]

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 1 candidates examined (receipt bdf454fc34b1).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  117. Definition · supporting · Population variance and sample variance

    Not formalized
    The resulting estimator is biased, however, and is known as the biased sample variation.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 8ce78da3260e).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  118. Definition · main statement · Population variance and sample variance › Population variance

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    In general, the population variance of a finite population of size ⁠ N ⁠ with values xi is given by σ2= 1N∑Ni = 1 (xi − μ)2 = 1N∑Ni = 1(x2i − 2μxi + μ2)= (1N∑Ni = 1 x2i) − 2μ(1N∑Ni = 1 xi) + μ2= E[x2i] − μ2, where the population mean is μ = E[xi] = 1N∑Ni = 1 xi and ⁠ E[x2i] = (1N∑Ni = 1 x2i) ⁠, where E is the expectation value operator.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance more general

    def ProbabilityTheory.variance {Ω : Type*} {mΩ : MeasurableSpace Ω} : (Ω → ℝ) → MeasureTheory.Measure Ω → ℝ

    Mathlib.Probability.Moments.Variance · line 64

    Why more general: Lean takes the variance against an arbitrary measure, and the passage is its reading when each of the N entries carries the same weight.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 2eb6542c9550 · pair key 856513fe1ad7.

    ProbabilityTheory.variance_eq_sub more general

    theorem ProbabilityTheory.variance_eq_sub {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsProbabilityMeasure μ] {X : Ω → ℝ} : MeasureTheory.MemLp X 2 μ → ProbabilityTheory.variance X μ = ∫ (x : Ω), (X ^ 2) x ∂μ - (∫ (x : Ω), X x ∂μ) ^ 2

    Mathlib.Probability.Moments.Variance · lines 226–228

    Why more general: Lean states the identity for any probability measure with a square-integrable variable, where the passage's mean is that variable's integral.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 46b9042cc8a8 · pair key 18ebf6f39fb7.

    ProbabilityTheory.variance_eq_integral more general

    theorem ProbabilityTheory.variance_eq_integral {Ω : Type*} {mΩ : MeasurableSpace Ω} {X : Ω → ℝ} {μ : MeasureTheory.Measure Ω} : AEMeasurable X μ → ProbabilityTheory.variance X μ = ∫ (ω : Ω), (X ω - ∫ (x : Ω), X x ∂μ) ^ 2 ∂μ

    Mathlib.Probability.Moments.Variance · lines 157–159

    Why more general: Lean gives the variance as the integral of the squared deviation from the mean against any measure, and equal weights on N entries turn that integral into the passage's average.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key f68064d9b658 · pair key 20b393eacc02.
  119. Theorem · supporting · Population variance and sample variance › Population variance

    Not formalized
    The population variance can also be computed using[13] σ2 = 1N2∑i < j(xi − xj)2 = 12N2∑Ni, j = 1(xi − xj)2.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 7e2797c5dd36).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  120. Theorem · supporting · Population variance and sample variance › Population variance

    Not formalized
    (The right side has duplicate terms in the sum while the middle side has only unique terms to sum.) This is true because 12N2∑Ni, j = 1 (xi − xj)2=12N2∑Ni, j = 1(x2i − 2xixj + x2j)=12N∑Nj = 1(1N∑Ni = 1 x2i) − (1N∑Ni = 1 xi)(1N∑Nj = 1 xj) + 12N∑Ni = 1(1N∑Nj = 1 x2j)=12(σ2 + μ2) − μ2 + 12(σ2 + μ2)=σ2.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt a579b47c6fc8).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  121. Theorem · supporting · Population variance and sample variance › Population variance

    Formalized

    Formalized: Mathlib states a more general result (proved).

    The population variance matches the variance of the generating probability distribution.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.HasLaw.variance_eq more general

    theorem ProbabilityTheory.HasLaw.variance_eq {Ω : Type*} {mΩ : MeasurableSpace Ω} {P : MeasureTheory.Measure Ω} {μ : MeasureTheory.Measure ℝ} {X : Ω → ℝ} : ProbabilityTheory.HasLaw X μ P → ProbabilityTheory.variance X P = ProbabilityTheory.variance id μ

    Mathlib.Probability.HasLaw · lines 203–205

    Why more general: The declaration is stated for an arbitrary real random variable on an arbitrary measure space, so it reaches well beyond the finite population the passage treats.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 5fed0466373d · pair key b49ce312c721.

    ProbabilityTheory.variance_id_map more general

    theorem ProbabilityTheory.variance_id_map {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {X : Ω → ℝ} : AEMeasurable X μ → ProbabilityTheory.variance id (MeasureTheory.Measure.map X μ) = ProbabilityTheory.variance X μ

    Mathlib.Probability.Moments.Variance · lines 338–339

    Why more general: It is stated for any measure and any real random variable, where the passage has one finite population with equal weights.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 0dff05111f45 · pair key 792d1a53fbaa.
  122. Definition · main statement · Population variance and sample variance › Sample variance › Biased sample variance

    Not formalized

    With community edits: Not formalized. As published it is Partial.

    This is generally referred to as sample variance or empirical variance.

    Show in article · Read it in context on Wikipedia

  123. Definition · main statement · Population variance and sample variance › Sample variance › Biased sample variance

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    We take a sample with replacement of ⁠ n ⁠ values Y1, ..., Yn from the population of size ⁠ N ⁠, where n < N, and estimate the variance on the basis of this sample.[15] Directly taking the variance of the sample data gives the average of the squared deviations:[16] S̃2Y = 1n∑ni = 1(Yi − Ȳ)2 = (1n∑ni = 1 Y2i) − Ȳ2 = 1n2∑i, j  :  i < j(Yi − Yj)2. (See the section § Population variance for the derivation of this formula.) Here, Ȳ denotes the sample mean: Ȳ = 1n∑ni = 1 Yi.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_eq_sub states part

    theorem ProbabilityTheory.variance_eq_sub {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsProbabilityMeasure μ] {X : Ω → ℝ} : MeasureTheory.MemLp X 2 μ → ProbabilityTheory.variance X μ = ∫ (x : Ω), (X ^ 2) x ∂μ - (∫ (x : Ω), X x ∂μ) ^ 2

    Mathlib.Probability.Moments.Variance · lines 226–228

    Community Agent The pipeline said more general; changed by Meta FAIR 2 (agent) — owned by Jack McCarthy on 9 Nov 2026.

    Why only part: Under the uniform measure on the n draws, ProbabilityTheory.variance_eq_sub gives the passage's middle equality: the average squared deviation equals the mean of the squares minus the squared sample mean. It does not state the passage's last equality, which writes the same quantity as 1/n^2 times the sum of (Y_i - Y_j)^2 over pairs i < j. So it formalizes one clause of the passage, not all of it.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 1ac510d79630 · pair key 733157251e78.

    ProbabilityTheory.variance more general

    def ProbabilityTheory.variance {Ω : Type*} {mΩ : MeasurableSpace Ω} : (Ω → ℝ) → MeasureTheory.Measure Ω → ℝ

    Mathlib.Probability.Moments.Variance · line 64

    Why more general: The library defines the variance against an arbitrary measure on an arbitrary space, where the passage fixes the uniform measure on n draws.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 6e1843bda8c2 · pair key 085509a16405.
  124. Notation · supporting · Population variance and sample variance › Sample variance › Biased sample variance

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    …S̃2Y = 1n∑ni = 1(Yi − Ȳ)2 = (1n∑ni = 1 Y2i) − Ȳ2 = 1n2∑i, j  :  i < j(Yi − Yj)2. (See the section § Population variance for the derivation of this formula.) Here, Ȳ denotes the sample mean: Ȳ = 1n∑ni = 1 Yi.

    Show in article · Read it in context on Wikipedia

    Fintype.expect_eq_sum_div_card more general

    theorem Fintype.expect_eq_sum_div_card {ι : Type*} {K : Type*} [Semifield K] [CharZero K] [Fintype ι] (f : ι → K) : (Finset.univ.expect fun i ↦ f i) = (∑ i, f i) / ↑(Fintype.card ι)

    Mathlib.Algebra.BigOperators.Expect · lines 371–372

    Why more general: It is stated over any characteristic-zero semifield where the passage has the real observations of one sample.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 34dbb0ab082a · pair key 57fb6fdc4d4c.

    Finset.expect more general

    def Finset.expect {ι : Type*} {M : Type*} [AddCommMonoid M] [_root_.Module ℚ≥0 M] : Finset ι → (ι → M) → M

    Mathlib.Algebra.BigOperators.Expect · lines 60–61

    Why more general: The library defines the average over an arbitrary finset of a module over the nonnegative rationals, which includes the passage's real sample.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 4cce67d7686c · pair key dc54fb0e33b2.

    TauCeti.finiteMean_eq_inv_smul_sum more general

    theorem TauCeti.finiteMean_eq_inv_smul_sum {R : Type*} {E : Type*} {ι : Type*} [Field R] [AddCommGroup E] [_root_.Module R E] [Fintype ι] (z : ι → E) : TauCeti.finiteMean R z = (↑(Fintype.card ι))⁻¹ • ∑ i, z i

    TauCeti.Analysis.Convex.FiniteMean · lines 56–58

    Why more general: The declaration is stated for a family of vectors in any module over any field, where the passage has n real observations.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 9d44fa3c609e · pair key 6e250bef9fc3.

    TauCeti.finiteMean more general

    def TauCeti.finiteMean (R : Type*) {E : Type*} {ι : Type*} [Field R] [AddCommGroup E] [_root_.Module R E] [Fintype ι] : (ι → E) → E

    TauCeti.Analysis.Convex.FiniteMean · lines 50–51

    Why more general: It defines the equal-weight mean of a finite family of vectors over any field, which covers the passage's real sample.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 4 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 5eb945411a80 · pair key a3bf1dd5d452.
  125. Theorem · supporting · Population variance and sample variance › Sample variance › Biased sample variance

    Not formalized
    Since the Yi are selected randomly, both Ȳ and S̃2Y are random variables.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt 36af7d0539ec).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  126. Theorem · main statement · Population variance and sample variance › Sample variance › Biased sample variance

    Not formalized
    Their expected values can be evaluated by averaging over the ensemble of all possible samples {Yi} of size ⁠ n ⁠ from the population. For S̃2Y this gives: E[S̃2Y]= E[1n∑ni = 1 (Yi − 1n∑nj = 1 Yj)2]= 1n∑ni = 1 E[Y2i − 2nYi∑nj = 1 Yj + 1n2∑nj = 1 Yj∑nk = 1 Yk]= 1n∑ni = 1(E[Y2i] − 2n(∑j ≠ i E[YiYj] + E[Y2i]) + 1n2∑nj = 1 ∑nk ≠ j E[YjYk] + 1n2∑nj = 1 E[Y2j])= 1n∑ni = 1(n − 2nE[Y2i] − 2n∑j ≠ i E[YiYj] + 1n2∑nj = 1 ∑nk ≠ j E[YjYk] + 1n2∑nj = 1 E[Y2j])= 1n∑ni = 1[n − 2n(σ2 + μ2) − 2n(n − 1)μ2 + 1n2n(n − 1)μ2 + 1n(σ2 + μ2)]= n − 1nσ2.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 4 candidates examined (receipt c0dde8ebc63b).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  127. Theorem · supporting · Population variance and sample variance › Sample variance › Biased sample variance

    Formalized

    Formalized by 2 parts together (more general):

    1. ProbabilityTheory.variance_eq_sub Mathlib (proved)
      • the population variance is the second moment of Y_i minus the square of the mean
    2. ProbabilityTheory.IndepFun.integral_fun_mul_eq_mul_integral Mathlib (proved)
      • for independent Y_i and Y_j the expectation of their product is the product of their means, which is mu squared
    Here σ2 = E[Y2i] − μ2 derived in the section is population variance and E[YiYj] = E[Yi]E[Yj] = μ2 due to independency of Yi and ⁠ Yj ⁠.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance_eq_sub states part

    theorem ProbabilityTheory.variance_eq_sub {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} [MeasureTheory.IsProbabilityMeasure μ] {X : Ω → ℝ} : MeasureTheory.MemLp X 2 μ → ProbabilityTheory.variance X μ = ∫ (x : Ω), (X ^ 2) x ∂μ - (∫ (x : Ω), X x ∂μ) ^ 2

    Mathlib.Probability.Moments.Variance · lines 226–228

    Why only part: Mathlib states this for any real random variable on a probability space and writes both expectations as integrals against the measure.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 87164c945c09 · pair key 6c9619962039.

    ProbabilityTheory.IndepFun.integral_fun_mul_eq_mul_integral states part

    theorem ProbabilityTheory.IndepFun.integral_fun_mul_eq_mul_integral {Ω : Type*} {𝕜 : Type*} [RCLike 𝕜] {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {X Y : Ω → 𝕜} : ProbabilityTheory.IndepFun X Y μ → MeasureTheory.AEStronglyMeasurable X μ → MeasureTheory.AEStronglyMeasurable Y μ → ∫ (ω : Ω), X ω * Y ω ∂μ = (∫ (x : Ω), X x ∂μ) * ∫ (x : Ω), Y x ∂μ

    Mathlib.Probability.Independence.Integration · lines 423–426

    Why only part: Mathlib states it for functions into any field satisfying RCLike, with the expectation written as an integral.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key a9ddb9e6c960 · pair key acfa13b8e515.
  128. Theorem · main statement · Population variance and sample variance › Sample variance › Biased sample variance

    Not formalized
    Hence S̃2Y gives an estimate of the population variance σ2 that is biased by a factor of n − 1n because the expectation value of S̃2Y is smaller than the population variance (true variance) by that factor.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 82e8cc6f150f).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  129. Definition · main statement · Population variance and sample variance › Sample variance › Biased sample variance

    Formalized

    Formalized: Batteries states this up to a convention difference (a definition).

    For this reason, S̃2Y is referred to as the biased sample variance.

    Show in article · Read it in context on Wikipedia

    Batteries.RunningStats.variance close match

    def Batteries.RunningStats.variance : Batteries.RunningStats → Float

    Batteries.Data.RunningStats · lines 60–61

    Why close, not exact: Batteries carries the sample as Float values in a running accumulator of count, mean and accumulated squared deviations, and divides the accumulated sum by the count.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 4453b52.
    • Location key 78bc788d364a · pair key ccc7aa83e87a.
  130. Definition · main statement · Population variance and sample variance › Sample variance › Unbiased sample variance

    Formalized

    Formalized: Batteries states this up to a convention difference (a definition).

    Correcting for this bias yields the unbiased sample variance, denoted ⁠ S2 ⁠: S2 = nn − 1S̃2Y = nn − 1[1n∑ni = 1(Yi − Ȳ)2] = 1n − 1∑ni = 1(Yi − Ȳ)2

    Show in article · Read it in context on Wikipedia

    Batteries.RunningStats.sampleVariance close match

    def Batteries.RunningStats.sampleVariance : Batteries.RunningStats → Float

    Batteries.Data.RunningStats · lines 65–66

    Why close, not exact: Batteries carries the sample as Float values in a running accumulator and divides the accumulated squared deviations by one less than the count, which is the same quantity as the scaled estimator of the passage.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 4453b52.
    • Location key eea33060c3d2 · pair key 4aa354425f03.

    MetaMathlibExt.initialSegmentSampleVariance special case

    def MetaMathlibExt.initialSegmentSampleVariance : (ℕ → ℕ) → ℕ → ℝ

    MathlibExt.Analysis.Asymptotics.TaylorInitialSampleVariance · line 41

    Why a special case: Atlas defines it only for sequences of natural numbers, so it is the passage's estimator at samples whose values are whole numbers.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key f45f80d4407c · pair key 0bdc08734942.
  131. Definition · supporting · Population variance and sample variance › Sample variance › Unbiased sample variance

    Not formalized
    The use of the term n − 1 is called Bessel's correction, and it is also used in sample covariance and the sample standard deviation (the square root of variance).

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 0 candidates examined (receipt cc2c5556ccb2).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  132. Theorem · supporting · Population variance and sample variance › Sample variance › Unbiased sample variance

    Partial

    Partial: Mathlib states part of this (proved).

    The square root is a concave function and thus introduces negative bias (by Jensen's inequality), which depends on the distribution, and thus the corrected sample standard deviation (using Bessel's correction) is biased.

    Show in article · Read it in context on Wikipedia

    Real.strictConcaveOn_sqrt states part

    theorem Real.strictConcaveOn_sqrt : StrictConcaveOn ℝ (Set.Ici 0) fun x ↦ √x

    Mathlib.Analysis.Convex.SpecificFunctions.Pow · lines 90–92

    Why only part: Mathlib records strict concavity of the square root on the nonnegative half-line, with the domain written as the set of reals at least zero.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 3b39393bff29 · pair key 259420b924ae.

    ConcaveOn.le_map_integral states part

    theorem ConcaveOn.le_map_integral {α : Type*} {E : Type*} {m0 : MeasurableSpace α} [NormedAddCommGroup E] [NormedSpace ℝ E] [CompleteSpace E] {μ : MeasureTheory.Measure α} {s : Set E} {f : α → E} {g : E → ℝ} [MeasureTheory.IsProbabilityMeasure μ] : ConcaveOn ℝ s g → ContinuousOn g s → IsClosed s → (∀ᵐ (x : α) ∂μ, f x ∈ s) → MeasureTheory.Integrable f μ → MeasureTheory.Integrable (g ∘ f) μ → ∫ (x : α), g (f x) ∂μ ≤ g (∫ (x : α), f x ∂μ)

    Mathlib.Analysis.Convex.Integral · lines 205–208

    Why only part: Mathlib states Jensen's inequality for a concave real-valued map on a closed set, leaving the size of the gap unquantified as the passage does.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 7b07724bf663 · pair key 2f2187538c5b.
  133. Theorem · supporting · Population variance and sample variance › Sample variance › Unbiased sample variance

    Not formalized
    The square root is a concave function and thus introduces negative bias (by Jensen's inequality), which depends on the distribution, and thus the corrected sample standard deviation (using Bessel's correction) is biased.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 1 candidates examined (receipt 63af6d1c82f5).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  134. Theorem · supporting · Population variance and sample variance › Sample variance › Unbiased sample variance

    Not formalized
    The unbiased estimation of standard deviation is a technically involved problem, though for the normal distribution using the term n − 1.5 yields an almost unbiased estimator.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 1 candidates examined (receipt c01a0c6f8814).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  135. Theorem · supporting · Population variance and sample variance › Sample variance › Unbiased sample variance

    Not formalized
    The unbiased sample variance is a U-statistic for the function f(y1, y2) = (y1 − y2)2/2, meaning that it is obtained by averaging a 2-sample statistic over 2-element subsets of the population.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 772c63d315d0).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  136. Theorem · supporting · Population variance and sample variance › Sample variance › Unbiased sample variance › Example

    Not formalized
    For a set of numbers {10, 15, 30, 45, 57, 52, 63, 72, 81, 93, 102, 105}, if this set is the whole data population for some measurement, then variance is the population variance 932.743 as the sum of the squared deviations about the mean of this set, divided by 12 as the number of the set members.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 75e4b1133359).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  137. Theorem · supporting · Population variance and sample variance › Sample variance › Unbiased sample variance › Example

    Not formalized
    If the set is a sample from the whole population, then the unbiased sample variance can be calculated as 1017.538 that is the sum of the squared deviations about the mean of the sample, divided by 11 instead of 12.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 0995bfd6f951).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  138. Theorem · supporting · Population variance and sample variance › Sample variance › Distribution of the sample variance

    Not formalized
    Being a function of random variables, the sample variance is itself a random variable, and it is natural to study its distribution.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 1 candidates examined (receipt 2e3d8d55159f).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  139. Theorem · main statement · Population variance and sample variance › Sample variance › Distribution of the sample variance

    Not formalized
    In the case that Yi are independent observations from a normal distribution, Cochran's theorem shows that the unbiased sample variance S2 follows a scaled chi-squared distribution (see also: asymptotic properties and an elementary proof):[17] (n − 1)S2σ2 ∼ χ2n − 1, where σ2 is the population variance.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt 33345b980123).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  140. Theorem · supporting · Population variance and sample variance › Sample variance › Distribution of the sample variance

    Not formalized
    As a direct consequence, it follows that E(S2) = E(σ2n − 1χ2n − 1) = σ2, and[18] Var[S2] = Var(σ2n − 1χ2n − 1) = σ4(n − 1)2Var(χ2n − 1) = 2σ4n − 1.…

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 96e3cdce1077).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  141. Theorem · supporting · Population variance and sample variance › Sample variance › Distribution of the sample variance

    Not formalized
    As a direct consequence, it follows that E(S2) = E(σ2n − 1χ2n − 1) = σ2, and[18] Var[S2] = Var(σ2n − 1χ2n − 1) = σ4(n − 1)2Var(χ2n − 1) = 2σ4n − 1.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 3ca0a1efbfe3).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  142. Theorem · main statement · Population variance and sample variance › Sample variance › Distribution of the sample variance

    Not formalized
    If Yi are independent and identically distributed, but not necessarily normally distributed, then[19] E[S2] = σ2;  Var[S2] = σ4n(κ − 1 + 2n − 1) = 1n(μ4 − n − 3n − 1σ4), where κ is the kurtosis of the distribution and μ4 is the fourth central moment.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt a1b06c9c8fc9).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  143. Theorem · main statement · Population variance and sample variance › Sample variance › Distribution of the sample variance

    Not formalized
    If the conditions of the law of large numbers hold for the squared observations, S2 is a consistent estimator of σ2.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt 0b5a0cc6339c).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  144. Theorem · supporting · Population variance and sample variance › Sample variance › Distribution of the sample variance

    Not formalized
    One can see indeed that the variance of the estimator tends asymptotically to zero.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt c6f94c48a5ad).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  145. Theorem · main statement · Population variance and sample variance › Sample variance › Samuelson's inequality

    Not formalized
    Values must lie within the limits ⁠ ȳ ± σY(n − 1)1/2 ⁠.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt 878e1679a060).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  146. Theorem · main statement · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    When a single new observation xn + 1 is added to a set of n observations with mean x̄n and variance s2n, the new variance s2n + 1 can be expressed using a recursive updating formula. Based on the identity for the sum of squares provided by Chan et al. (1983):[24] ns2n + 1 = (n − 1)s2n + nn + 1(xn + 1 − x̄n)2

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt 4b1f70d61e0f).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  147. Theorem · main statement · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    If xn + 1 = x̄n ± sn√n + 1n, then the variance will remain unchanged.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt 4ab81d30413f).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  148. Theorem · main statement · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    Accordingly, if the new observation is closer to the mean ( |xn + 1| < x̄n + sn√n + 1n), then the variance will decrease - and if is further from the mean ( |xn + 1| > x̄n + sn√n + 1n), then the variance will increase.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt a5946bece73f).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  149. Theorem · supporting · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    Using the updating formula for the sum of squares ( SS): SSn + 1 = SSn + nn + 1(xn + 1 − x̄n)2

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt 183954c48da2).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  150. Theorem · supporting · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    Substituting the relationship for sample variance ( SS = (n − 1)s2): ns2n + 1 = (n − 1)s2n + nn + 1(xn + 1 − x̄n)2

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 70db8555ecb6).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  151. Theorem · supporting · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    Setting s2n + 1 = s2n: ns2n = (n − 1)s2n + nn + 1(xn + 1 − x̄n)2

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt e9a61e226c89).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  152. Theorem · supporting · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    ns2n = (n − 1)s2n + nn + 1(xn + 1 − x̄n)2
    s2n = nn + 1(xn + 1 − x̄n)2

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 994474f7b7ba).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  153. Theorem · supporting · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    Solving for xn + 1 yields: xn + 1 = x̄n ± sn√n + 1n

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt b0a37b0a873a).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  154. Theorem · supporting · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    For population variance ( σ2 = SSn), the updating formula is: (n + 1)σ2n + 1 = nσ2n + nn + 1(xn + 1 − μn)2

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 8b7cc8b7910b).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  155. Theorem · supporting · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    Setting σ2n + 1 = σ2n: (n + 1)σ2n = nσ2n + nn + 1(xn + 1 − μn)2

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt d34921506689).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  156. Theorem · supporting · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    (n + 1)σ2n = nσ2n + nn + 1(xn + 1 − μn)2
    σ2n = nn + 1(xn + 1 − μn)2

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 45c74cba827b).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  157. Theorem · supporting · Population variance and sample variance › Sample variance › Effect of adding one observation on variance

    Not formalized
    Solving for xn + 1 yields: xn + 1 = μn ± σn√n + 1n

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt b147cdeae858).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  158. Theorem · main statement · Population variance and sample variance › Relations with the harmonic and arithmetic means

    Not formalized
    It has been shown[25] that for a sample {yi} of positive real numbers, σ2y ≤ 2ymax(A − H), where ymax is the maximum of the sample, ⁠ A ⁠ is the arithmetic mean, ⁠ H ⁠ is the harmonic mean of the sample and σ2y is the (biased) variance of the sample.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 0a1186c388dc).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  159. Theorem · main statement · Population variance and sample variance › Relations with the harmonic and arithmetic means

    Not formalized
    This bound has been improved, and it is known that variance is bounded by σ2y≤ ymax(A − H)(ymax − A)ymax − H,σ2y≥ ymin(A − H)(A − ymin)H − ymin, where ymin is the minimum of the sample.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt da4386e13eb7).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  160. Theorem · supporting · Moment of inertia

    Not formalized
    The moment of inertia of a cloud of n points with a covariance matrix of Σ is given by[citation needed] I = n(13 × 3tr(Σ) − Σ).

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt b35bed5ff187).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  161. Theorem · supporting · Moment of inertia

    Not formalized
    That is, there is the most variance in the x direction. Physicists would consider this to have a low moment about the x axis so the moment-of-inertia tensor is I = n[0.200010.100010.1].

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt b521a1b9fd78).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  162. Definition · main statement · Semivariance

    Not formalized
    The semivariance is calculated in the same manner as the variance but only those observations that fall below the mean are included in the calculation: Semivariance = 1n∑i : xi < μ (xi − μ)2 It is also described as a specific measure in different fields of application.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 9e3491d6aa27).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  163. Theorem · supporting · Etymology

    Formalized

    Formalized: Mathlib states this up to a convention difference (proved).

    When there are two independent causes of variability capable of producing in an otherwise uniform population distributions with standard deviations σ1 and σ2, it is found that the distribution, when both causes act together, has a standard deviation √σ21 + σ22.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.IndepFun.variance_add close match

    theorem ProbabilityTheory.IndepFun.variance_add {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {X Y : Ω → ℝ} : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.IndepFun X Y μ → ProbabilityTheory.variance (X + Y) μ = ProbabilityTheory.variance X μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 408–415

    Why close, not exact: Lean measures the dispersion of each cause by its variance where the quote measures the same dispersion by its square root, and states the additivity in that encoding.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key b1d3523999b2 · pair key 4a8c30111836.

    ProbabilityTheory.IndepFun.variance_fun_add close match

    theorem ProbabilityTheory.IndepFun.variance_fun_add {Ω : Type*} {mΩ : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {X Y : Ω → ℝ} : MeasureTheory.MemLp X 2 μ → MeasureTheory.MemLp Y 2 μ → ProbabilityTheory.IndepFun X Y μ → ProbabilityTheory.variance (fun ω ↦ X ω + Y ω) μ = ProbabilityTheory.variance X μ + ProbabilityTheory.variance Y μ

    Mathlib.Probability.Moments.Variance · lines 418–420

    Why close, not exact: This is the pointwise form of the same additivity, again recording each spread as a variance rather than as its square root.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 1b28eca679a3 · pair key b1a4c9dba41f.
  164. Definition · main statement · Generalizations › For complex variables

    Partial

    Partial: Mathlib states a special case (a definition).

    If x is a scalar complex-valued random variable, with values in ⁠ ℂ ⁠, then its variance is ⁠ E[(x − μ)(x − μ)∗] ⁠, where x∗ is the complex conjugate of ⁠ x ⁠.

    Show in article · Read it in context on Wikipedia

    ProbabilityTheory.variance special case

    def ProbabilityTheory.variance {Ω : Type*} {mΩ : MeasurableSpace Ω} : (Ω → ℝ) → MeasureTheory.Measure Ω → ℝ

    Mathlib.Probability.Moments.Variance · line 64

    Why a special case: Mathlib introduces this quantity only for random variables with real values, the case of the quote in which the conjugated factor coincides with the first factor.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 1afa00dc5c31 · pair key 7ab82a8a7137.

    ProbabilityTheory.evariance special case

    def ProbabilityTheory.evariance {Ω : Type*} {mΩ : MeasurableSpace Ω} : (Ω → ℝ) → MeasureTheory.Measure Ω → ENNReal

    Mathlib.Probability.Moments.Variance · line 58

    Why a special case: This is the extended-nonnegative-real form of the same construction, again only for random variables with real values.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, low confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 07936add6fef · pair key d72f1a564601.
  165. Theorem · supporting · Generalizations › For complex variables

    Not formalized
    This variance is a real scalar.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt 9030ce753110).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  166. Definition · main statement · Generalizations › For vector-valued random variables › As a matrix

    Formalized

    Formalized: ATLAS (atlas-lean) states this up to a convention difference (a definition).

    If X is a vector-valued random variable, with values in ℝn, and thought of as a column vector, then a natural generalization of variance is E[(X − μ)(X − μ)T], where μ = E(X) and XT is the transpose of ⁠ X ⁠, and so is a row vector.

    Show in article · Read it in context on Wikipedia

    MathlibExt.Probability.IsotropicConstant.covarianceMatrix close match

    def MathlibExt.Probability.IsotropicConstant.covarianceMatrix {n : ℕ} : MeasureTheory.Measure (EuclideanSpace ℝ (Fin n)) → Matrix (Fin n) (Fin n) ℝ

    MathlibExt.Probability.Geometry.IsotropicConstant · line 35

    Why close, not exact: The library feeds the construction the law of the vector as a measure and reads the matrix off the covariances of the coordinate variables instead of writing one expectation of an outer product.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 5b7471cc37cb · pair key b0baef5183ef.

    TauCeti.covMatrix more general

    def TauCeti.covMatrix {ι : Type*} : MeasureTheory.Measure (EuclideanSpace ℝ ι) → Matrix ι ι ℝ

    TauCeti.Probability.Moments.Covariance · lines 52–53

    Why more general: The library builds the matrix for an arbitrary index type, so it also covers index sets that are not finite, and it takes the law of the vector as its input.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 43e9c0e24a66 · pair key a34dde8b0819.

    ProbabilityTheory.covarianceBilin more general

    def ProbabilityTheory.covarianceBilin {E : Type*} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [MeasurableSpace E] [BorelSpace E] : MeasureTheory.Measure E → E →L[ℝ] E →L[ℝ] ℝ

    Mathlib.Probability.Moments.CovarianceBilin · lines 51–53

    Why more general: Mathlib records the covariance of a measure as a continuous bilinear form on any real inner product space, including spaces of infinite dimension, rather than as a matrix.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 02207002e11d · pair key 4331c0c32c36.
  167. Theorem · main statement · Generalizations › For vector-valued random variables › As a matrix

    Formalized

    Formalized: TauCeti states a more general result (proved).

    The result is a positive semi-definite square matrix, commonly referred to as the variance-covariance matrix (or simply as the covariance matrix).

    Show in article · Read it in context on Wikipedia

    TauCeti.posSemidef_covMatrix more general

    theorem TauCeti.posSemidef_covMatrix {ι : Type*} [Fintype ι] (μ : MeasureTheory.Measure (EuclideanSpace ℝ ι)) [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp id 2 μ → (TauCeti.covMatrix μ).PosSemidef

    TauCeti.Probability.Moments.Covariance · lines 99–107

    Why more general: The statement is made for an arbitrary finite index type, and squareness is carried by the Lean type of the matrix rather than proved.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key da4415d4179d · pair key 9c528e6d7b33.

    ProbabilityTheory.isPosSemidef_covarianceBilin more general

    theorem ProbabilityTheory.isPosSemidef_covarianceBilin {E : Type*} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [MeasurableSpace E] [BorelSpace E] {μ : MeasureTheory.Measure E} : (ProbabilityTheory.covarianceBilin μ).toBilinForm.IsPosSemidef

    Mathlib.Probability.Moments.CovarianceBilin · lines 106–109

    Why more general: Mathlib phrases the covariance of a measure as a bilinear form on any real inner product space and calls that form positive semidefinite; the matrix shape is the coordinate reading of the same form.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key abe521219652 · pair key 2f97973acdc8.

    MathlibExt.Probability.IsotropicConstant.covarianceMatrix_posSemidef special case

    theorem MathlibExt.Probability.IsotropicConstant.covarianceMatrix_posSemidef {n : ℕ} (μ : MathlibExt.Probability.IsotropicConstant.LogConcaveLaw n) : (MathlibExt.Probability.IsotropicConstant.covarianceMatrix μ.law).PosSemidef

    MathlibExt.Probability.Geometry.IsotropicConstant · line 66

    Why a special case: Log-concavity of the law is assumed, and some of the passage's vectors (Gaussian or uniform on a convex body) meet it.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key d4f3aeb28256 · pair key c472ff50433c.
  168. Definition · main statement · Generalizations › For vector-valued random variables › As a matrix

    Partial

    Partial: TauCeti states a special case (a definition).

    If X is a vector- and complex-valued random variable, with values in ⁠ ℂn ⁠, then the covariance matrix is ⁠ E[(X − μ)(X − μ)†] ⁠, where X† is the conjugate transpose of ⁠ X ⁠.[citation needed] This matrix is also positive semi-definite and square.

    Show in article · Read it in context on Wikipedia

    TauCeti.covMatrix special case

    def TauCeti.covMatrix {ι : Type*} : MeasureTheory.Measure (EuclideanSpace ℝ ι) → Matrix ι ι ℝ

    TauCeti.Probability.Moments.Covariance · lines 52–53

    Why a special case: The library only treats vectors with real coordinates, where the matrix is real; the genuinely complex case, with the conjugate transpose acting, is not covered.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 259adf1d8f12 · pair key c5bf58fa8b0c.
  169. Theorem · supporting · Generalizations › For vector-valued random variables › As a matrix

    Partial

    Partial: TauCeti states a special case (proved).

    …with values in ⁠ ℂn ⁠, then the covariance matrix is ⁠ E[(X − μ)(X − μ)†] ⁠, where X† is the conjugate transpose of ⁠ X ⁠.[citation needed] This matrix is also positive semi-definite and square.

    Show in article · Read it in context on Wikipedia

    TauCeti.posSemidef_covMatrix special case

    theorem TauCeti.posSemidef_covMatrix {ι : Type*} [Fintype ι] (μ : MeasureTheory.Measure (EuclideanSpace ℝ ι)) [MeasureTheory.IsFiniteMeasure μ] : MeasureTheory.MemLp id 2 μ → (TauCeti.covMatrix μ).PosSemidef

    TauCeti.Probability.Moments.Covariance · lines 99–107

    Why a special case: Only the real-coordinate vectors are covered, and square-integrability of the law is assumed.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 7b425b6da7ce · pair key 1d884d786b81.
  170. Definition · main statement · Generalizations › For vector-valued random variables › As a scalar

    Not formalized
    Another generalization of variance for vector-valued random variables X, which results in a scalar value rather than in a matrix, is the generalized variance ⁠ det(C) ⁠, the determinant of the covariance matrix.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 4 candidates examined (receipt b3aba428e26d).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  171. Theorem · main statement · Generalizations › For vector-valued random variables › As a scalar

    Not formalized
    This results in E[(X − μ)T(X − μ)] = tr(C), which is the trace of the covariance matrix.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 4 candidates examined (receipt 60401b5ce5a1).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.

Data: view model (JSON) · Literature Links records (JSONL)