Code review request for issue 89

gcode-mirror / ai-contest

Automatically exported from code.google.com/p/ai-contest

0 stars 1 forks source link

Code review request for issue 89 #94

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago

This branch patches the process_ratings.py script to adjust the recorded Elo 
score by the confidence in estimating it. The end result is that scores would 
eventually converge on Elo, but gradually rise from the bottom rather than 
jumping around randomly like they currently do. This should fix issues like a 
new bot losing its first game and jumping *up* near the top rather than being 
near the bottom.

Original issue reported on code.google.com by Jake.McA...@gmail.com on 10 Sep 2010 at 4:19

GoogleCodeExporter commented 9 years ago

Set review issue status to: Started

Original comment by cameron.jp@gmail.com on 10 Sep 2010 at 4:54

Changed state: Started

GoogleCodeExporter commented 9 years ago

Here is how it changes a more recent version of the ranking (with lots of new 
bots). I include number of games played to show how it tends to change the rank 
of the bot.

http://hpaste.org/fastcgi/hpaste.fcgi/view?id=29882#a29882

Personally I find the change agreeable.

Original comment by Jake.McA...@gmail.com on 11 Sep 2010 at 12:02