§1, p. 759, "It is clear that" — P(S_n^j ≥ −x, max_{i≠j}X_i < x) ≥ P(S_n^j ≥ −x) − P(max_{i≠j}X_i ≥ x)
ProvedNagaevLD.LowerBound.clear_steplarge-deviationslower-boundp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be real random variables on a probability space , let , let be real and an index. Then
This is the elementary inequality applied to and . It reduces the probability on the right-hand side of the bound on to two one-sided tail probabilities, each of which is controlled by Chebyshev's inequality.
Formalization Note The page prints "" on the left and "" on the right, while (1.57) and the bound on use . The statement here uses on the left and therefore on the right, which is the form the argument needs. The event is written "there is with ".
Preamble
import Mathlib import Definitions.Def_NagaevLD_LowerBound_Setting open MeasureTheory ProbabilityTheory
Formal statement
namespace NagaevLD.LowerBound
/-- §1, p. 759, "It is clear that": `P(S_n^j ≥ −x, max_{i≠j} X_i < x) ≥ P(S_n^j ≥ −x) − P(max_{i≠j} X_i ≥ x)`
(the page prints `≤ x` and `> x`; this is the form the chain uses with (1.57)'s `< x`). -/
theorem clear_step {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(n : ℕ) (X : Fin n → Ω → ℝ) (hXm : ∀ i, Measurable (X i)) (x : ℝ) (j : Fin n) :
P.real {ω | -x ≤ Sj n X j ω} - P.real {ω | ∃ i, i ≠ j ∧ x ≤ X i ω} ≤
P.real {ω | -x ≤ Sj n X j ω ∧ ∀ i, i ≠ j → X i ω < x} := by sorry
end NagaevLD.LowerBound
Source
Nagaev, Large deviations of sums of independent random variables, Ann. Probab. 7 (1979), p. 759, §1, "It is clear that" (before (1.58))
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.