Originally Posted by F Paulsson
This is the problem with your reasoning; you can't extrapolate this.
Since my direct approach didn't work, I'll try an analogy.
If, at the end of each month, I count the number of wounded US soldiers in Iraq, I might end up with, say 5%. This might lead me to extrapolate that only 5% of all soldiers are wounded in Iraq. This is not true. The wounded soldiers go home and are replaced with unwounded ones.
Unless someone can show me that people who quit poker are equally likely to be net winners and net losers, the statistic of how many players are winners in PokerTracker is, if not meaningless, close to it.
Well I agree with you, but as a counter to this, I play pretty regularly and see a good amount of players. The ones who play more are most certainly the winners. I just did a filter of players I had 1k+ hands with and on average they win at a rate of 1.08 PTBB/100. The average winrate for everyone (includes the winners above) is -7.08 PTBB/100. I still can't figure out how it's so low (since without rake it'd be even and rake is only 2-4 PTBB/100 depending on stakes), but that's just the number the database gives me.
But the point is that I'm observing the winners more, meaning that my database will be more likely to be more accurate because it's seeing this trend that you point out in this post. I still have about a 65-35 distribution and I really don't believe that is true.
I think it has more to do with the fact that people move up, as in the Irexes thread, to the point where they lose. So you have someone who can beat up to 25nl. They'll beat 25nl, move up to 50nl, and then continue to try to beat 50nl losing their entire BR in the process. They may have a good run that convinces them they're good enough to beat it and that 25nl is beneath them. Then you have the people who beat 50nl and can't beat 100nl, etc. on up to the $200/$400 NLHE games where people keep taking shots.
If we follow the axiom that a player will move up the point where they lose, the only winning players in the world would be the ones beating the highest stakes games. Since obviously some people can swallow their pride and realize they're making good money where they are and there's no point in moving up, that's why there are more than 5 winning players on each site. I think I've seen numbers ranging from 5-10% and that's what I'd guess is accurate.
I also think that FP has a point for smaller samples, because you basically see the winners as those who get lucky, and mine's probably skewed because I table-select to avoid the very players who would be winning in my db over the long-term so representatively I have more hands from the weaker opponents than from the stronger ones. But I think the factor above is much more influential than just that losers quit. I think when you combine the factors that winners play more, losers quit faster, and that most winners will move up to the point where they don't win anymore, that's where you get the 5-10% figure.