Battle of the Crowns -- calibration runs

[Dann Corbit]

Geschrieben von: / Posted by: Dann Corbit at 25 April 2000 00:38:44:
Here are the current battle of the crowns calibration runs. Particularly compelling was the 50 hour battle this weekend between Amy and Amnon. After 50 games, it was dead even.
Date Time Size File
-------- ------ ------ ------------------
04/20/00 12:32p 9,800 Amy-Amyan.pgn
04/24/00 12:43p 45,296 Amy-Anmon.pgn
04/18/00 08:49p 18,797 Amy-Arasan.pgn
04/17/00 01:04p 5,277 Amy-Averno.pgn
04/17/00 01:05p 8,605 Amy-Bringer.pgn
04/17/00 01:05p 13,210 Amy-Crafty.pgn
04/19/00 03:40p 8,799 Amy-GnuChess5.pgn
04/17/00 01:05p 7,472 Amy-Knightx.pgn
04/20/00 12:44p 15,390 Amyan-Anmon.pgn
04/21/00 03:01p 11,187 Amyan-Ant.pgn
04/21/00 03:00p 10,235 Amyan-Averno.pgn
04/20/00 12:44p 13,355 Amyan-Bringer.pgn
04/24/00 03:29p 39,901 Amyan-TSCP.pgn
04/24/00 03:30p 39,595 Amyan-Zchess.pgn
04/17/00 01:05p 11,445 Anmon-Bringer.pgn
04/17/00 01:05p 8,615 Arasan-Ant.pgn
04/17/00 01:05p 54,587 Arasan-Bringer.pgn
04/18/00 05:29p 8,019 Arasan-Sos.pgn
04/18/00 08:46p 10,263 Arasan-Ssechess.pgn
04/19/00 03:36p 10,338 Arasan-TCBishop.pgn
04/18/00 05:29p 39,644 Arasan-Zchess.pgn
04/19/00 03:35p 8,865 Averno-Bringer.pgn
04/17/00 01:05p 4,415 Averno-Monik.pgn
04/17/00 01:05p 1,828 Averno-Pierre.pgn[1]
04/17/00 01:05p 10,779 Averno-Zchess.pgn
04/17/00 01:05p 36,903 Comet-LGoliath.pgn
04/17/00 01:05p 40,030 Comet-Sos.pgn
04/17/00 01:05p 12,187 Comet-TCBishop.pgn
04/17/00 01:05p 8,333 Crafty-TCBishop.pgn
04/17/00 01:05p 9,309 Dchess-Zchess.pgn
04/17/00 01:05p 3,319 Dragon-Zchess.pgn
04/17/00 01:05p 11,081 Skaki-Tscp.pgn
04/17/00 01:05p 10,012 Ssechess-Zchess.pgn
04/17/00 01:05p 8,054 Tscp-Zchess.pgn
[1] Pierre did not respond correctly, and so lost all games on forfeits. I will have to examine the setup carefully.


My FTP site

[Jon Dart]

Geschrieben von: / Posted by: Jon Dart at 25 April 2000 20:01:07:
Als Antwort auf: / As an answer to: Battle of the Crowns -- calibration runs geschrieben von: / posted by: Dann Corbit at 25 April 2000 00:38:44:

Here are the current battle of the crowns calibration runs. Particularly compelling was the 50 hour battle this weekend between Amy and Amnon. After 50 games, it was dead even.

Interesting data. I also see you did a long match between Bringer
and Arasan. By the way, the last game in this match is missing
a result. Also some of the opening classifications are wrong.
E.g. the last game in this file is not an Albin Counter Gambit.
And the round 4 game is misclassified as a Reverse Grob - it
looks like a Queen's Gambit.
--Jon

[Dann Corbit]

Geschrieben von: / Posted by: Dann Corbit at 25 April 2000 20:10:26:
Als Antwort auf: / As an answer to: Re: Battle of the Crowns -- calibration runs geschrieben von: / posted by: Jon Dart at 25 April 2000 20:01:07:

Here are the current battle of the crowns calibration runs. Particularly compelling was the 50 hour battle this weekend between Amy and Amnon. After 50 games, it was dead even.

Interesting data. I also see you did a long match between Bringer
and Arasan.
By the way, the last game in this match is missing
a result.
Also some of the opening classifications are wrong.
E.g. the last game in this file is not an Albin Counter Gambit.
And the round 4 game is misclassified as a Reverse Grob - it
looks like a Queen's Gambit.

As you know, Arasan is one of my favorite programs. I suggest it to anyone who is intimidated about trying to set up a Winboard program.
This is true for a number of the matches. This experiment is not part of the tournament, per se. The only purpose is to find an approximate ELO of the players so that accurate seedings can be formed. Since the round robins will be held within 4 distinct groups [collected by ability], a bad seeding would render fragments of the experiment as useless. Hence, I start a batch of games going over the night or over the weekend for a set of machines. Whenever someone needs to use the machine in the morning, they just stop the program. Hence, the number of games played may not be an even number and many games will be stopped in mid-stream. Since partial games don't add anything, I usually cut them off the end of the pile, but sometimes I will probably forget.
After two months, I will have enough one hour games to get an accurate ELO assessment. This data will be used for the sole purpose of creation of seeding pairs for the Battle of the Crowns tournament.
I'm not terribly surprised by this. The classification is not done by hand, but instead by the program Extract by David Barnes. I have collected classifications from several sources and there are undoubtably bugs in my classification collection.


My FTP site