Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Chess program similarity experiment (Results)

Author: Will Singleton

Date: 11:13:23 01/26/99

Go up one level in this thread



Results of Bionic epd test --

So far, looking at the results of Ferret, EXchess and Amateur,
it's striking how similar they are (looking at Bionic's moves,
not the opponents).  With the exception of round 5, the scores
are very close.  Should be interesting to see more results.


Amateur 0.9h, Mac G3/300
--------------------------

Bionic --
Round   Total   Matched Pct

1        60     33   55%
2        38     11   29%
3        30     23   77%
4        40     27   68%
5        17      9   53%
6        44     29   66%
7        53     30   57%
8        47     26   55%
9        32     24   75%
10       22     17   77%
11       22     14   64%
Total   405    243   60%

Opponent --
Round   Total   Matched Pct  Opponent

1        60     16   27%      The King
2        38     12   32%      Nimzo99
3        30     16   53%      Alexs
4        39     16   41%      Ant
5        17      9   53%      Diep
6        43     22   51%      Kallisto II
7        54     27   50%      CilkChess
8        47     21   45%      Arthur
9        31     20   65%      BugChess
10       21      9   43%      Morphy 3.0
11       22      7   32%      Delta
Total   402    175   44%


I noticed that in many of the positions, Amateur had the
"key" move earlier in the search, with a similar eval.  So
the results would change a lot if the criterion for a match
changed (ie, +- 0.10 eval points, or different time to move).

Will





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.