Proposition 3.1 and the admissible parameter in Theorem 3.3 #
The cubic inverse is unique on [0,2] for target footrule in [-1/2,0]. Uniqueness among all real roots would be false, as checked below.
Proposition 3.1, evaluated from the source's exact integral definitions.
theorem
Papers.Rockel2026XiFootrule.relaxed_coefficients_endpoints :
Verification.relaxedFootrule 0 = 0 ∧ Verification.relaxedFootrule 2 = -1 / 2 ∧ Verification.relaxedXi 0 = 0 ∧ Verification.relaxedXi 2 = 12 * Real.log 2 - 8
The source's phrase "unique real solution" needs the admissible interval restriction.
theorem
Papers.Rockel2026XiFootrule.weighted_lower_bound_closed
(C : ProbabilityTheory.Copula 2)
(μ : ℝ)
(hμ : μ ∈ Set.Icc 0 2)
:
Theorem 3.2 with explicitly evaluated coefficients.
theorem
Papers.Rockel2026XiFootrule.negative_footrule_lower_bound
(C : ProbabilityTheory.Copula 2)
(hy : C.spearmanFootrule ∈ Set.Icc (-1 / 2) 0)
:
∃! μ : ℝ, μ ∈ Set.Icc 0 2 ∧ Verification.footruleCubic C.spearmanFootrule μ = 0 ∧ Verification.xiClosed (Verification.relaxedParameter μ) ≤ C.chatterjeeXi
Theorem 3.3's lower estimate at every nonpositive attainable footrule value.