Theorem 6, p. 14 — I-ELF is strictly incentive compatible under belief independence
ProvedForecastComp.IELF.theorem_6Let and , and let be a strictly proper scoring rule with values in , measurable in the report, and with values in on reports in when (the requirement of the I-ELF mechanism, p. 14). Then the Independent-Event Lotteries mechanism is strictly incentive compatible under belief independence: for every forecaster , every belief-independent joint law of the outcomes and the others' reports with , and every with ,
Together with Theorem 7 this is the paper's main result: I-ELF rewards truthful reporting even though only one forecaster is selected.
Formalization Note Belief independence is encoded as a product of per-event joint laws (Definition 8). Measurability of in the report makes the selection probabilities well-defined expectations.
import Mathlib import Definitions.Def_ForecastComp_IELF_Model
namespace ForecastComp.IELF
/-- Theorem 6, p. 14 (proof p. 34): I-ELF is strictly incentive compatible under belief
independence for `m ≥ 1` events, for `R` a strictly proper scoring rule with values in `[0,1]`
(in `[0,1)` when `m ≥ 2`, as the mechanism requires on p. 14). -/
theorem theorem_6 {n : ℕ} (hn : 2 ≤ n) (R : ℝ → Bool → ℝ)
(hprop : IsStrictlyProper R) (hbdd : IsUnitBounded R)
(hmeas : ∀ x, Measurable (fun y => R y x))
(m : ℕ) (hm : 1 ≤ m) (hlt : 2 ≤ m → ∀ y ∈ Set.Icc (0:ℝ) 1, ∀ x, R y x < 1) :
StrictlyICUnderBeliefIndep (ielfSelProb (m := m) (n := n) R) := by sorry
end ForecastComp.IELF
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.