A uniform quantitative improvement of eta <= 2 xi #
theorem
Verification.cdf_centered_quantitative_gap
(C : ProbabilityTheory.Copula 2)
(t : ↑unitInterval)
:
21 / 20 * (∫ (u : ↑unitInterval), C.conditionalCDF t u - ↑u) ^ 2 ≤ (∫ (u : ↑unitInterval), (C.conditionalCDF t u - ↑u) ^ 2) + 1 / 1000
The endpoint bounds of a CDF force a positive square-error contribution.
This estimate separates every xi=1/4 copula uniformly from eta=1/2.