Last month I introduced the Convolution Integral (CI) as a measure of Elo-gaining skills ONLY, as one of the complaints about the Test Against Pure Luck (TAPL) is that it measured all kinds of skill, including negative skill. The explanation and formula for the CI are in parts vi and vii of the comment on my profile page.

After tracking the CI’s of the Top 25’er’s for a month, I have come to the following tentative conclusions:

1. An unassisted CI significantly higher than 2.00 is probably not sustainable over a significant number of games. All the CI’s over 2 in the top 25 either a) played a relatively small number of games, or b) have been fingered as gpa’ers in the recent posts about cabals, or c) both.

2. A player with a CI between 1.00 and 2.00 has enough strength to advance to the next level of table when it becomes available.

3. A player with a CI between -0.50 and 1.00 is probably at the peak of his or her game. Even with a mildly negative CI, a player can maintain standing because of the boost given to the elo scores in the lower part of the point range of a table. A CI close to zero does not preclude a player from maintaining standing through tough slogging, such as is shown by canarioz and mikeypoo.

4. A player with a CI between -1.00 and -0.50 can maintain standing but needs improvement in his game. My personal monthly CI’s are in this range because I am too bull-headed about wanting to attack the #1 player in the game.

I will run the stats on this particular version of the CI for another month. I want to see how the rule changes affect play among the top 25’ers. After this month, I may recompute the formula for the CI based on top 25’er data rather than the current data, which came from contestants in May’s lifetime TAPL competition.