We haven't won yet: latest (10/19) cliff's notes on AP scandal

[admiralfluff]

Guys - it is much easier to use a results based detection criteria for fixed limit hands. Even a short sample can be conclusive proof of anamolous play. For example, if you have a 100 hand sample of a player in your DB, and they are winning at several hundred BB/100 over that sample, it is extremely likely something fishy is going on (dumping or cheating). With NL, it is more difficult to draw these types of conclusions, because it could just be the result of several stackings. I am working on some scripts to detect anamolous patterns based on several derived parameters. I'm looking for some help from decent windows coders and stats guys.