Two-step zero Coin probability at adapted stopping times
ProvedCountingLLL.ReciprocalTest.suppressed_stopping_coin_zeroformalizationprobability
At any history-adapted optional stopping time for the fully suppressed execution, conditional in joint-probability form on reaching a specified Coin-oracle state, the probability of a zero Coin return in the next two atomic steps is exactly B/2. The stopping time may be infinite, and the identity is valid for every Coin context.
Preamble
import Definitions.Def_CountingLLL_ReciprocalSuppressedTransition open MeasureTheory ProbabilityTheory CountingLLL.ReciprocalTest open scoped ENNReal
Formal statement
theorem CountingLLL.ReciprocalTest.suppressed_stopping_coin_zero (B : ℝ≥0∞) (hB : B ≤ 1) (T : Entropy → Option ℕ)
(H : (n : ℕ) → (Fin (n+1) → State) → Prop)
(hT : ∀ ω n, T ω = some n ↔ H n (fun j => allSuppressedRun ω j.val))
(s : State) (k : Context) (hs : s.phase = .coinOracle k) :
let u : State := {s with oracleUsed := s.oracleUsed+1, phase := .coinFair k}
let v : State := coinReturn k false {u with fairUsed := u.fairUsed+1}
entropyLaw B {ω | ∃ n, T ω = some n ∧ allSuppressedRun ω n = s ∧
allSuppressedRun ω (n+1) = u ∧ allSuppressedRun ω (n+2) = v} =
entropyLaw B {ω | ∃ n, T ω = some n ∧ allSuppressedRun ω n = s} * (B/2) := by sorrySource
Liu, Wang, Yin, Zhang, Zhou, A Counting and Sampling Lovász Local Lemma, STOC submission #442, §6.2, Algorithm 6 and Lemma 6.5; adaptive progress ingredient for the contracted walk.