[Nomad84]

I'm not sure what data can be exported from PT to Excel (either directly or via database queries), but it might be possible to get a rough idea of the percentage of winning players if we could export # of hands, win rate (or net winnings), and standard deviation. If necessary, we could approximate SD. We could then determine how confident we are that each player is an overall winner pretty easily. By converting that to a percentage likelyhood that each is a winner and averaging this column, we could get an approximate percentage of overall long term winners, I think. It may be necessary to only include players with some minimum number of hands, since we won't be very confident in any results from players without many hands. By that, I mean that if someone has relatively few hands, we may be only 60% sure that they are winners, even if they have a great win rate. I would expect a disproportionate number of players with relatively few hands in most databases. This would tend to cause our estimate to be too high since even the losers (with few hands) will still be 35% likely to be winners, for example. I'm not sure what the hand cutoff should be though. If we require too many hands, we will exclude players who lose quickly and give up, while including players who go on a sustained rush even though they are long-term losers. I doubt this method could provide a very good estimate, but it may be better than just throwing out educated guesses. I don't have the motivation to actually give it a shot. Besides, my database is relatively small. However, if someone else wanted to try it, it should be a fairly quick process and it would be interesting to see the results.