| |
BGonline.org Forums
ELO
Posted By: Christian Plenz In Response To: ELO (Jake Jacobs)
Date: Sunday, 20 December 2009, at 6:42 a.m.
ELO is a good thing but maybe we are at a point where we should think of a new rating system. Most of the downsides of the ELO system in Backgammon are mentioned in the thread below
Our goal was and is to use the two Players ELO's as a determinator of the evaluations of a match. So why not speak directly in terms of Match Performances as a description of Rating for a player.
Lets say: Player A' Rating is 55%
Player B' Rating is 65%
Player A has performed 55% in his Matches, that means he has won 55% against average Opponents for example or 50% against better (over average performing) players (which had 55% Performance themselves)
Then we have A against B and we can make legitimate forecasts for the match. Player B has 60.31%.
One thing what imO is superior than with the ELO is that you can see 1) the percentage of matches won and lost 2) against the average opponent's level
For example: 60% Matches won against 45% Opp-Avg = 54% Performance-Rating
Every players performance is built of his result against different players of different strenghts (off course with a matching coefficient of the latest results)
What we have to figure out is to find the parameters for manipulations and adjustments in different Match lenghts and Player performances long and short time.
But I think if we can get this the result should probably be better than the ELO's cause we have predictions of the Truths instead of deviated numbers who were trying to be true.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.