By: Marcel Goh

In reply to Marcel Goh.

Elaborating on my earlier comment, letting and be conditionally independent trials of , I think I can show that assuming Lemma A.2 as well as (24) and (25) from Theorem 3.1 in the older paper. This is not exactly (28), and I’m not sure how to transform the into a , since gives us a lot more information than from the coupling $X_1+Y_1 = X_2+Y_2$. (At least I think it does.)