| |
BGonline.org Forums
More swiss tournaments, best in three shorter matches and clocks
Posted By: Maik Stiebler In Response To: More swiss tournaments, best in three shorter matches and clocks (Matt Cohn-Geier)
Date: Saturday, 11 October 2008, at 9:07 a.m.
Matt, your analysis seems to be in line with mine (look at the JT bit - I used the table with the 50 elo skill difference. The long match - 21 points in my case - appears to have about 75% more skill content than the set of one-pointers, about the same as in your analysis).
However, your conclusion hinges on the assumption that the Jacobs-Trice tables are accurate. I doubt that, which is of course not meant as a stab against "Can a Fish Taste Twice As Good". It's just that the tables were probably not meant to be used in this context, and accuracy of the predicted skill content in matches of different length was therefore probably not a design goal. The one piece of evidence that I have is that the JT tables say that one-pointers have less skill content than the plain FIBS elo formula predicts, whereas conventional wisdom has it that the plain FIBS elo formula understates the skill content in one-point matches.
Under certain simple assumptions (perfect cube efficiency, fixed and equal gammon rate for both sides, fixed game winning percentage of the stronger player, I think) the set of one-pointers beats longer matches hands down. That's the "MET method" section in the post I linked to. The answer is probably somewhere between those poles, and depends on the specific players.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.