Actual BV Lévy origin dyadic deficit slopes converge to zero almost surely
ProvedAvramDividend.Classical.bv_actual_origin_dyadic_deficit_zero_ae_ws01calmost-sure-convergenceavram-classicalbounded-variationdyadic-zero-slopews01c
For the actual bounded-variation spectrally negative Lévy process, the entire sequence of dyadic residual slopes d−(X_(2^(-n))−X_0)/2^(-n) converges almost surely to zero. This is the correct Cauchy upgrade: the source-complete Fubini-stationarity theorem gives an almost-sure Cauchy sequence, the actual scaled-tail theorem gives convergence in measure, and the already Proved Cauchy-plus-in-measure limit theorem yields the full almost-sure zero limit. This avoids the false inference of almost-sure convergence from convergence in probability alone.
Preamble
import Mathlib import Theorems.Thm_AvramDividend_Classical_bv_actual_origin_dyadic_cauchy_ae_ws01c import Theorems.Thm_AvramDividend_Classical_bv_actual_dyadic_slope_tendstoInMeasure_zero_ws01c import Theorems.Thm_AvramDividend_Classical_bv_ae_cauchy_plus_in_measure_bundle_ws01c open MeasureTheory Filter Topology AvramDividend.Classical open scoped NNReal
Formal statement
theorem AvramDividend.Classical.bv_actual_origin_dyadic_deficit_zero_ae_ws01c
{Ω : Type*} [mΩ : MeasurableSpace Ω] {P : Measure Ω}
{ℱ : Filtration ℝ≥0 mΩ}
(X : SpectrallyNegativeLevy P ℱ) (hBV : X.BoundedVariation) :
∀ᵐ ω ∂P,
Tendsto (fun n : ℕ =>
X.drift -
(X.X (((1 : ℝ≥0) / 2) ^ n) ω - X.X 0 ω) /
(((1 : ℝ≥0) / 2) ^ n : ℝ))
atTop (𝓝 (0 : ℝ)) := by sorrySource
Source-complete WS01C origin zero-slope formalisation assembled from work/avram_ws01c_origin_zero_slope_full_assembly_draft.lean. Invokes existing canonical Proved Cauchy-plus-measure theorem eb4d928e, plus two newly submitted specific actual-process intermediate targets.