1-30-06 [poker]I just discovered a new stat which appears to be the single most correlated simple stat for fishyness. I call it "PFWtSD%". This is similar to the standard Poker Tracker stat "WtSD%" , which is "went to show down percentage", eg. how many hands do they take to showdown after seeing a flop. The "PF" in mine means from "pre flop", which means they take a hand all the way from being dealt to show down. For a solid typical TAG type player, PFWtSD% is usually in the 2-5% range. That comes from folding a lot preflop or postflop, but also from betting postflop and getting others to fold so there's no showdown. An uber fish will have a PFWtSD% over 10%. I've seen them above 40% (!!). This means they're not betting much, and are calling bets, and when they are betting they're getting called a lot. Unlike other stats, the correlation here is fantastic. For example, if you just use "vpip" (voluntarily put in pot), you can find the loose preflop players, but some of them are very tight/aggressive postflop, so they aren't necessarilly fishies, but even those guys have a low PFWtSD, since they will usually fold or get their opponent to fold before showdown. If they're showing down a lot (high PFWtSD%) and their big pot % is high, then they're a money factory.
It's a funny truth that these statistics are far more reliably than their winnings. Obviously over a large # of hands, you can just look at someone's winnings and tell if they're giving away money or taking in money, but that number is incredibly variable. Players that I'm sure are good show big losses over 5k hands. Players I know are terrible show wins. The actual money won is incredibly affected by short term luck, which makes it a bad way to pick fishies. Someone's play style stats are also affected by short term cards, but much less so.