Proof of Proposition 5, p. 26 β at the true mean R, the capital process π¦α΅’(R) is a martingale for Ο(Lβ, β¦, Lα΅’)
ProvedRiskControl.WSR.K_martingaleLet be a probability measure on concentrated on , let be its mean, let and , and let the sample have the product law , so that are i.i.d. with law . Let be the capital process of Proposition 5 and (see the definition RiskControl.WSR.Process). Then the process
is a martingale with respect to under ; in particular, for ,
This is the martingale property at the heart of the proof of Proposition 5: since is -measurable and , betting against the true mean does not change the expected capital.
Formalization Note is the law of the loss for the fixed , which lies in by the standing assumption of Β§3.1; the losses of the i.i.d. calibration points are then i.i.d. with law . The paper's martingale is indexed by with trivial and ; it is extended constantly past (index , with for ), so that it is a process indexed by , as Mathlib's Martingale requires. The nonnegativity half of the paper's sentence ("non-negative martingale") is the separate item RiskControl.WSR.factor_nonneg. No condition on is needed: is predictable for every real .
import Mathlib import Definitions.Def_RiskControl_WSR_Process open MeasureTheory
namespace RiskControl.WSR
/-- Proof of Proposition 5, p. 26 (arXiv:2101.02703v3): at the true mean `R = β« x dQ` of the
`[0, 1]`-valued losses, the capital process `π¦_i = π¦_i(R; Ξ»)`, `i = 0, β¦, n`, is a martingale
for the natural filtration `β±_i = Ο(L_1, β¦, L_i)` under the i.i.d. law `QβΏ`; it is extended
constantly past `n` (index `min i n`) to a process indexed by `β`. -/
theorem K_martingale (Q : Measure β) [IsProbabilityMeasure Q]
(hQ : βα΅ x βQ, x β Set.Icc (0 : β) 1) (n : β) (Ξ΄ : β) :
Martingale (fun i (Ο : Fin n β β) => K n Ξ΄ Ο (min i n) (β« x, x βQ)) (filt n)
(Measure.pi fun _ : Fin n => Q) := by sorry
end RiskControl.WSR
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.