Drawdown singleton masses tend to zero at large scale
ProvedKyprianou.Ch08.StageB.p2m81_drawdown_singleton_tendsto_zero_publickyprianou-ch08p2m81public-bridgesingletonsweak-convergence
For the drawdown probability chosen at scale d, its Laplace transform at rate one tends to one as d tends to infinity. Markov's inequality applied to one minus exp(-Y) then bounds any fixed positive singleton mass by the Laplace deficit divided by one minus exp(-x), so the singleton masses tend to zero.
Preamble
import Mathlib open MeasureTheory ProbabilityTheory Filter Set Topology open scoped NNReal ENNReal set_option autoImplicit false
Formal statement
theorem Kyprianou.Ch08.StageB.p2m81_drawdown_singleton_tendsto_zero_public
(f : ℝ → ℝ) (hpos : ∀ s : ℝ, 0 < s → 0 < f s)
(hdraw : ∀ d : ℝ, 0 < d → ∃ ρ : Measure ℝ,
IsProbabilityMeasure ρ ∧ ρ (Iio (0 : ℝ)) = 0 ∧
∀ s : ℝ, 0 < s →
(f d - f s) * (∫⁻ y : ℝ, ENNReal.ofReal (Real.exp (-s * y)) ∂ρ).toReal =
f d * (1 - s / d))
(x : ℝ) (hx : 0 < x) :
Tendsto
(fun d : ℝ =>
((if hd : 0 < d then Classical.choose (hdraw d hd) else 0) : Measure ℝ) ({x} : Set ℝ))
atTop (nhds (0 : ENNReal)) := by
sorry
Source
Kyprianou (2006), Chapter 8, Theorem 8.1. Public helper for the geometric descending potential, drawdown singleton limit.