ATCSecure has responded NUMEROUS times that by rev 2 with multi-path, this is a non-issue. Why should he waste any more time on a pattern matcher when this exact problem won't be a problem any more?
So...Chaeplin is guessing and getting at least part of it right because there is a "single path" (and not too many variables) ...but when there is a "multi path" this will all be mixed up too much?
All he could show was the mixer, not the source and destination, he could not link the transactions...