Author: Norm Pollock
Date: 06:19:18 02/18/05
Go up one level in this thread
On February 18, 2005 at 07:58:14, Steve Coladonato wrote: >Thanks Norm. I guess I wasn't looking for a "Book" file. > >One other statistic that might be useful would be the delta between the avg ELO >for white vs black in each of your groupings. e.g. +20 would indicate that >white had an average ELO rating 20 points higher than black and -20 would >indicate that white had an average ELO rating 20 points lower than black. > >Regards. > >Steve Hi Steve, Definitely. It would be great if I could show that the average elo for white was equal to the average elo for black. And if not, be able to state the difference (delta). Unfortunately that statistic is not available as far as I know, with the database manager(Scid). A spot check of just a few players with high elo in my second sample (2450-2650 elo) shows that they had black more often than white. That would cause the white/black difference to contract. Conclusion: The 7% white/black difference of this sample should be interpreted as a 6%-8% range because of the possibility that on average White had a significantly lower/higher elo than Black. But as I pointed out in the study, I deliberately chose a sample with a narrow elo range to minimize this problem. -Norm
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.