The central limit theorem
ProvedFamousTheorems.tendstoInDistribution_cltThe central limit theorem.
Let be independent, identically distributed real random variables with finite second moment, mean and variance . Then
The remarkable content is universality: the limit does not remember the distribution of at all, only its mean and variance. A sum of many small independent contributions is Gaussian whether the summands are coin flips, dice, or anything else with finite variance. This is why the normal distribution appears everywhere in statistics and physics, and it is the reason a scaling is the right one — any other power gives either or divergence.
De Moivre proved the case of fair coin flips in 1733; Laplace extended it to general binomial trials. The modern general statement with necessary and sufficient conditions is due to Lindeberg (1920), Lévy and Feller. Along with the law of large numbers it is one of the two pillars of classical probability: the law of large numbers says the average converges, and the central limit theorem describes the size and shape of the error.
Formalization note. TendstoInDistribution is weak convergence of the laws; HasLaw Y (gaussianReal 0 v) P'
says is a centred Gaussian of variance under , and MemLp (X 0) 2 P is the finite-second-moment
hypothesis. The result is Mathlib's ProbabilityTheory.tendstoInDistribution_inv_sqrt_mul_sum_sub.
import Mathlib
namespace FamousTheorems
open MeasureTheory ProbabilityTheory Filter
open scoped Real Topology
theorem tendstoInDistribution_clt {Ω Ω' : Type*} {mΩ : MeasurableSpace Ω}
{mΩ' : MeasurableSpace Ω'} {P : Measure Ω} {P' : Measure Ω'} {X : ℕ → Ω → ℝ} {Y : Ω' → ℝ}
[IsProbabilityMeasure P] [IsProbabilityMeasure P']
(hY : HasLaw Y (gaussianReal 0 (Var[X 0; P]).toNNReal) P')
(hX : MemLp (X 0) 2 P) (hindep : iIndepFun X P)
(hident : ∀ i : ℕ, IdentDistrib (X i) (X 0) P P) :
TendstoInDistribution
(fun (n : ℕ) ω ↦ (√n)⁻¹ * (∑ k ∈ Finset.range n, X k ω - n * P[X 0]))
atTop Y (fun _ ↦ P) P' := by sorry
end FamousTheorems