There are so many "what ifs" and plausible deniabilities here, which is why I think Samaurai is just presenting BS. Granted I haven't looked at Wasabi code, but doesn't this make an assumption that largest output is change when it might not be?
The transactions like this are (afaik) not used as a proof in front of a court, so plausible deniability doesn't matter much. The transactions are used to trace the user to a point where they make a mistake (e.g. send some of that change to KYC exchange). At that point the enforcement agency can with reasonable precision target that specific person and they are quite likely to find some actual evidence. The "peeling" behavior of Wasabi has been used in the same way in the previous Plus token and Kucoin hacks...
reply