Original canonical dividend discount mesh error is always finite and converges to zero pathwise
ProvedAvramDividend.Classical.avram_jump6_B_original_canonical_discount_error_finite_and_vanishingFor the actual original admissible dividend strategy, at any fixed nonnegative stopped physical clock and any nonnegative discount, the precise ENNReal coefficient q/(j+1) times the true dividendMeasure (Ioc 0 t) is never infinite and its toReal vanishes as j tends to infinity. Original rightLimit D interval mass formula ensures finiteness at every clock, without a finite jump-activity assumption. Supplies both strict-pre-ruin stopped cash error finiteness and passage to the real Stieltjes mesh limit.
Preamble
import Mathlib import Definitions.Def_AvramDividend_Classical_DividendStrategy import Theorems.Thm_AvramDividend_Classical_dividendMeasure_Ioc_rightLimit open MeasureTheory Set Filter Topology AvramDividend.Classical open scoped NNReal ENNReal noncomputable section
Formal statement
theorem AvramDividend.Classical.avram_jump6_B_original_canonical_discount_error_finite_and_vanishing
{Ω : Type*} [mΩ : MeasurableSpace Ω]
{𝓕 : Filtration ℝ≥0 mΩ}
(D : ℝ≥0 → Ω → ℝ) (hD : IsDividendStrategy 𝓕 D)
(ω : Ω) (t : ℝ≥0) (q : ℝ) (hq : 0 ≤ q) :
let E : ℕ → ℝ≥0∞ := fun j =>
ENNReal.ofReal (q * ((1 : ℝ) / ((j : ℝ) + 1))) *
dividendMeasure D ω (Ioc (0 : ℝ) (t : ℝ))
(∀ j, E j ≠ ⊤) ∧
Tendsto (fun j => (E j).toReal) atTop (𝓝 (0 : ℝ)) := by sorry