§1, p. 759 — P(S_n^j ≥ −x) ≥ 1 − B_n²/x² for centred summands
ProvedNagaevLD.LowerBound.chebyshev_Sjchebyshev-inequalitylarge-deviationslower-boundp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be independent real random variables on a probability space with and for every . Write , and . Then for every and every index
The sum of the summands other than has variance , so it falls below with probability at most . In Nagaev's lower bound this keeps the rest of the sum from cancelling a large summand .
Formalization Note The hypothesis of the page is stated as "each is square integrable" (Lean's MemLp (X i) 2 P); without it Mathlib's variance of a non-square-integrable variable is and the bound would be false.
Preamble
import Mathlib import Definitions.Def_NagaevLD_LowerBound_Setting open MeasureTheory ProbabilityTheory
Formal statement
namespace NagaevLD.LowerBound
/-- §1, p. 759: `P(S_n^j ≥ −x) ≥ 1 − B_n²/x²` for centred, square-integrable, independent summands. -/
theorem chebyshev_Sj {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(n : ℕ) (X : Fin n → Ω → ℝ) (hXm : ∀ i, Measurable (X i))
(hindep : iIndepFun (fun i => X i) P)
(hL2 : ∀ i, MemLp (X i) 2 P) (hmean : ∀ i, ∫ ω, X i ω ∂P = 0)
(x : ℝ) (hx : 0 < x) (j : Fin n) :
1 - NagaevLD.FukNagaev.Bn2 P X / x ^ 2 ≤ P.real {ω | -x ≤ Sj n X j ω} := by sorry
end NagaevLD.LowerBound
Source
Nagaev, Large deviations of sums of independent random variables, Ann. Probab. 7 (1979), p. 759, §1, "On the other hand, P(S_n^j ≥ −x) ≥ 1 − B_n²/x²"
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.