Theorem 7, p. 14 — I-ELF is limit accurate for independent events
ProvedForecastComp.IELF.theorem_7forecasting-competitionlimit-accuracyp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let and let be any scoring rule with values in on reports in . Then is limit accurate with respect to and the set of joint laws under which the event outcomes are independent: for every and every there is such that for all , all , all and every forecaster with for all ,
The threshold depends only on , and (and ), never on or the reports.
Formalization Note Independent outcome laws are parametrized by their marginals . Strict properness of and the requirement for are not assumed: the result holds for every -valued (a disclosed strengthening of the statement; the paper's satisfies it).
Preamble
import Mathlib import Definitions.Def_ForecastComp_IELF_Model
Formal statement
namespace ForecastComp.IELF
/-- Theorem 7, p. 14 (proof pp. 35–36): I-ELF is limit accurate for every scoring rule `R` with
values in `[0,1]` and all joint laws under which the event outcomes are independent. -/
theorem theorem_7 {n : ℕ} (hn : 2 ≤ n) (R : ℝ → Bool → ℝ) (hbdd : IsUnitBounded R) :
IsLimitAccurateIndep R (fun m θ y i => ielfSelProbIndep (m := m) (n := n) R θ y i) := by sorry
end ForecastComp.IELF
Source
Witkowski et al., arXiv:2101.01816v2, p. 14, Theorem 7 (proof pp. 35–36, Appendix M); Definition 7, p. 13
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.