This is an intersting question. I see two ways you could continue.
1) If you have some rating(s) of 'style' for each of the 600 players, you could use those ratings, or perhaps some interaction of them, as a covariate How exactly you would do this depends on what the ratings of style are like, and what your exact hypothesis is. 2) Lacking such ratings of style, you might try to derive some. Assuming that these 600 players are top grandmasters, then you MIGHT be able to do something with multidimensional scaling. What this would require is getting a bunch of people to rate how 'similar' different players are. The trouble is, with 600 people to rate, you'd need a lot of raters, and they'd have to be knowldegable enough to give sensible ratings of the players. Perhaps this is doable; you might try contacting a bunch of very knowldgable players and offering them money to do the job. HTH Peter Peter L. Flom, PhD Assistant Director, Statistics and Data Analysis Core Center for Drug Use and HIV Research National Development and Research Institutes 71 W. 23rd St www.peterflom.com New York, NY 10010 (212) 845-4485 (voice) (917) 438-0894 (fax) >>> [EMAIL PROTECTED] 2/27/2004 4:26:52 AM >>> Thanks for the responses guys. The ELO system does indeed give reasonably accurate estimates of the probability of winning. A simple logit model based on the difference in ELO rating will give a reasonable estimate of the probabilities of players winning. I suppose what I am asking, in a sense, is could this prediction be improved by somehow accounting for the head-to-head record of the two participants in addition to the ELO rating difference. A corollorary to that is obviously whether the head-to-head effect is significant of itself. I am not a statistician (I'm an engineer) but I am numerate and am prepared to study the appropriate area. However I don't know what that area is! What would a suitable null hypothesis be? The motivation for the question by the way comes from the idea that styles of play may suit certain opponents, e.g. my style of play might be relatively ineffective against many players but very effective against a small number of players, some of whom may be significantly better (in the sense they are ranked higher) than me. I have a large database of matches (about 13,000 for about 600 players) where I have the ratings and the outcomes of the games. What I am looking for is some statistical analysis I can do to begin addressing the above question (although obviously looking at head-to-head rather than the rather intangible style issue) . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . ================================================================= . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
