Setting up my HUD & PT3....

VPIP & PFR accumulate every hand, and therefore converge more quickly than almost all other stats. A sample as small as 20 hands will pretty much tell you what you need. With 50 hands on someone, the VPIP/PFR is reliable. After all, the difference between a 33/15 and 34/16 is pretty slim anyway.

The only time this is not solid is when someone get an extreme card distribution in their first 20 - 50 hands. Like getting a ton of great hands in a row... or the opposite, running card dead for the first few orbits. This quickly resolves itself and your Spidey senses will tell you that something is different about that villain until the stats catch up with them.

As MMM said don't worry about sample size when the stats are extreme - if someone is running 60/5 over 20 hands, they are loose and passive, if someone is running 4/2 over 20, they are a nit. The exceptions over even 20 or 30 hands to this will be rare. These are really the crucial types to differentiate quickly when you join a table, and it doesn't take many hands.

Note that a 60/40 or whatever is not necessarily a fish over 20 hands, since they have a more normal vpip/pfr ratio and could be someone running hot, wheras someone 60/5 is a sure bet to be a fish since they are calling and limping a ton.

For your more standard tags, differentiating say a 25/22 type and a nittier 18/15, yeah a hundred hands or so ought to provide a decent sample.

The ones to look out for with small sample sizes are stuff like a 30/15 over 30 hands may be a tight, aggressive player who's running hot, or a fish, so look for other evidence like posting blinds, sitting short, limping, minraising etc. Similarly a 25/10 could be a tag, or a card-dead fish, so again look for the signs that differentiate an aggressive vs a passive player.

I also tend to see maniacs take a little while to converge, like a 25/25 over 30 hands could end up being 50/50 over a hundred, they are just a little card dead again or something. With these players, I think you really do need to wait a little while to draw conclusions since a 25/25 who is 3betting and barrelling off a ton could just be a "decent" player who is running hot, so I'd play pretty straightforward with them until the sample size accumulated.

Originally Posted by BorisTheSpider

As MMM said don't worry about sample size when the stats are extreme - if someone is running 60/5 over 20 hands, they are loose and passive, if someone is running 4/2 over 20, they are a nit.

Ignoring the fact that it's impossible to run 4/2 over 20, I'll just say that someone so tight over 20 is not necessarily a supernit, as accurate tagg stats take much longer than loose fish stats to converge. a 5/5 over 20 could be 14/12 or 20/18, it's just too early to tell. All we can say is that they're almost certainly not a loose fish.

Originally Posted by BorisTheSpider

Note that a 60/40 or whatever is not necessarily a fish over 20 hands, since they have a more normal vpip/pfr ratio and could be someone running hot

60/40 over 20 could certainly be a 25/22 just running into good spots, but if you've got a decent sample at a stake, and know most regs -- I'd much rather assume that they're an aggressive fish, but if they've been playing that many hands you should have some other information to decide if they're leaning towards reg or fish.