The final inequality of Theorem 1.2 and its equality case #
For every stochastically increasing copula C,
ξ(C) ≤ 1 - (3/4)(1 - β(C))², with equality iff C = R_{β(C)}.
We also record ξ(R_b) = 1 - (3/4)(1 - b)², a consequence of the equality case of Lemma 5.2.
theorem
Papers.OrendayLaresRockel2026XiBeta.si_beta_mem_Icc
(C : ProbabilityTheory.Copula 2)
(hC : C.IsSI)
:
Stochastic increasingness forces β ∈ [0,1].
theorem
Papers.OrendayLaresRockel2026XiBeta.si_bound_data
(C : ProbabilityTheory.Copula 2)
(hC : C.IsSI)
:
∃ (s : MedianSection),
(∀ (u : ↑unitInterval), s.A ↑u = C.cdf ![u, ProbabilityTheory.Copula.unitHalf]) ∧ s.slopeData.q = (1 - C.blomqvistBeta) / 2 ∧ C.chatterjeeXi ≤ s.completion.chatterjeeXi ∧ s.completion.chatterjeeXi = 1 - 3 / 2 * s.slopeData.J ∧ 2 * s.slopeData.q ^ 2 ≤ s.slopeData.J
The two-branch bound, together with the data needed for its equality case.
theorem
Papers.OrendayLaresRockel2026XiBeta.xi_le_of_isSI
(C : ProbabilityTheory.Copula 2)
(hC : C.IsSI)
:
The final inequality of Theorem 1.2: ξ(C) ≤ 1 - (3/4)(1 - β(C))² for SI copulas.
theorem
Papers.OrendayLaresRockel2026XiBeta.xi_eq_iff_of_isSI
(C : ProbabilityTheory.Copula 2)
(hC : C.IsSI)
:
Equality case of Theorem 1.2: equality holds iff C = R_{β(C)}.