The December player stats appear not to have corrupted percentages. A glitch on the first day, where players stats were multiples of 100%, was fixed by Ryan. The glitch is automatically fixed for a player when that players plays at least one game after the fix.

The TAPL depends on accurate percentages stats. Weekly TAPL reports will resume when the top 100 players show enough game totals over 35 (Yate's Correction does not have to be used). Along with the TAPL, I will report on my new measure, the TAZD (Test Against Zero Datum).

I have also done some work on tweaking the Convolution Integral (CI). It measures a player's expected PPG according to what percentage firsts, seconds, etc. are in the player's profile and how many points each place is worth. The problem with the CI is that it does not pick up dom, so it is inaccurate for the 500 and 2000 level tables. I have been doing a least-square error analysis of the top 100 players' PPG's against a linear transformation of their CI's. Preliminary analysis suggests that the best linear fit of the CI is 96% place and 4% dom. The top 100 players' PPG's are roughly six times bigger than their CI's, which is consistent with the average level of table they play at being 600. Or you could say that the buy-in per game of the top 100 is roughly six times that of the buy-in at the zero level tables. The ratio of PPG to CI for a particular player can be used to measure that player's gutsiness, but gutsiness could be directly measured without having to calculate a CI if Ryan were to incorporate players' buy-ins into the stats *cough*.

The practical significance of the "96% place and 4% dom" is that 4th place would get 4% of the buy-in on average, instead of zero.