Author: KarinsDad
Date: 17:59:19 01/29/99
Go up one level in this thread
On January 29, 1999 at 20:22:57, James T. Walker wrote: >Hello Karinsdad, >It's 32 for 0-2099,24 for 2100-2399 and 16 for over 2400 (K). But the bottom >line as I see it is this. If your winning expectancy does not match the WE in >the formula guess what. They change your rating so that it does. So if you >continue to play someone 400 points above you and you don't eventually get your >win or draw in x number of games you are no longer 400 points below. Therefore >to stay 400 points below someone your WE must match the formula. It's true you >can't play even one game and lose and stay exactly 400 points below because by >virtue of the loss you drop 1 point and this continues untill you get your due >win. Then you are back where you started. At least I think that's the way it's >supposed to work. >Jim Walker Theoretically and for all intents and purposes, it does work this way at the +-300 point range. At 700 points difference, you would gain 32 points if you win, -1 points if you lose. The formula frequency of winning is about 55 loses and 1 win. The real frequency is probably about a 1000 to 1 (granted, I am totally guessing here, but I can never see myself beating a GM, there are other factors such as nerves, and he would always kick butt on my minor mistakes). So if you only played someone 700 points higher than you, you would gain 32 points on that one win and lose 55 points on the other 55 loses (if the formula is correct). If I am correct, you would gain 32 points on that one win, and lose 1000 points on the 1000 loses. Obviously, the system seems flawed, even if you do not believe as I do that the prediction formula is way off base at the greater ranges and is accurate. The reason that nobody notices it is that you only play someone 700 points higher than you once in a blue moon at an open, so losing the 1 point doesn't seem like a big deal and doesn't really hurt your rating. 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.