Author: KarinsDad
Date: 11:38:23 10/06/99
Go up one level in this thread
On October 06, 1999 at 14:07:34, Ricardo Gibert wrote: [snip] >> >>I think you are justifying a system that does not reflect playing strength as >>"accurately" as other systems. >> >>Although the system is consistent within itself, it IS less accurate than a >>system where the point is lost, but ratings adjustments are not made. Having a >>1200 player gain 32 points due to a 2400 player forfeiting to him (and vice >>versa) is kinda silly. Effectively what you have in the Canadian system is noise >>which distorts the accuracy (this being a relative term since no rating system >>is totally accurate) of the ratings. >> >>KarinsDad :) > >Distortions are swamped out by all the games that end normally. What percentage >of games does the average player win or lose by forfeit? Out of about maybe 75 tournament games, I have won I think 2 by forfeit. So, for me, it would be a little under 3%. It usually only happens in big money tournaments where someone loses a game and decides to quit the tournament without telling the TD. You are correct that the noise is very low. However, it is still there and still measurable. It would be interesting to take 100,000 Canadian games and re-calculate the ratings with and without the forfeiture rule. I doubt the greatest rating difference for anyone with a reasonable number of games would be much greater than about 5 points or so (unless the forfeiture was "recent", i.e. within the most recent games selected in the re-calculation), but I'm sure that it is enough to put a few people into a lower or higher section (class, or whatever the Canadian players call it). KarinsDad :)
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.