I wonder: should I use byo yomi for time keeping or initial time + incremental (like chess does)?
For now I've implemented byo yomi but I noticed that gnugo does not like that.
Search found 37 matches
- Thu Aug 24, 2023 9:00 am
- Forum: Go AI
- Topic: tournament manager
- Replies: 3
- Views: 502
perft
Hi, In chess we have "perft": count the number of valid moves upto a certain depth. It is used to validate the move-generator. I googled for it, but they're not known for Go programs? My program outputs the following for a board-size of 9: 1: 81 (81000.0 moves per second) 2: 6480 (925714.3...
- Wed Apr 05, 2023 6:53 pm
- Forum: Go AI
- Topic: tournament manager
- Replies: 3
- Views: 502
Re: tournament manager
Ok, I've added the following:
* seed board with random
* seed board with a board-layout from a random sgf in a sub-directory
* bugfixes
* re-licensed under the more liberal MIT license
* seed board with random
* seed board with a board-layout from a random sgf in a sub-directory
* bugfixes
* re-licensed under the more liberal MIT license
- Tue Apr 04, 2023 9:26 pm
- Forum: Go AI
- Topic: tournament manager
- Replies: 3
- Views: 502
tournament manager
Hi, I wrote a tournament manager for Go (GTP) programs ( https://github.com/folkertvanheusden/badank ). It works pretty well but I noticed that when running certain programs, it has only 25% unique games. E.g. 13090 games played, only 2998 unique sgf-files. In chess one uses an opening book to preve...
Re: checks
Hi Flok, Here are in general the techniques I use to ensure quality of my C++ software: googletest for unit testing: https://github.com/google/googletest gcov to measure test coverage: https://gcc.gnu.org/onlinedocs/gcc/Gcov.html run tests with sanitizers: https://github.com/google/sanitizers memor...
checks
Hi, Currently, when I've placed a stone, I regenerate the possible chains with for each chain a list of its liberties (I use that liberties-list to see if it is dead or not). After that, I do an assert if the all chains of the color that just placed the stone, if they all have 1 or more liberties. W...
- Thu Mar 30, 2023 3:39 pm
- Forum: Go AI
- Topic: References
- Replies: 2
- Views: 4732
Re: References
Archive of old mailing list: https://www.mail-archive.com/computer-go@computer-go.org/maillist.html GTP documentation http://www.lysator.liu.se/~gunnar/gtp/ CGOS http://www.yss-aya.com/cgos/ KGS GTP client: http://www.gokgs.com/download.jsp UEC Cup http://entcog.c.ooco.jp/entcog/new_uec/en/ Invite ...
- Wed Mar 15, 2023 6:45 pm
- Forum: Go AI
- Topic: New gogui release is coming
- Replies: 3
- Views: 890
Re: New gogui release is coming
The 1.5.2 release is online: https://github.com/Remi-Coulom/gogui Don't use Quaqua anymore, which should fix some crashes on macOS (issue #18). Support for large boards up to 52x52. Support for games with moving pieces such as chess or ataxx Drag-and-drop of sgf files (thanks to @zakki) Fix wrong p...
- Wed Mar 08, 2023 10:40 am
- Forum: Go AI
- Topic: Python version of CGOS by Zakki
- Replies: 19
- Views: 2124
Re: Python version of CGOS by Zakki
@zaki: any idea why my program 'DellaBaduckf6497eb' on cgos 9x9 is not matched against stop-0.9? It played 209 games sofar and none of them were against stop.
- Fri Mar 03, 2023 12:09 am
- Forum: Go AI
- Topic: Plenty of weak networks on CGOS 9x9
- Replies: 13
- Views: 1091
Re: Plenty of weak networks on CGOS 9x9
Yes. Currently I have for eval the end-score (including komi).Rémi Coulom wrote: ↑Tue Feb 28, 2023 5:56 pmWell, the idea can be useful in an alpha-beta search as well. If you give a bonus to a group with eyes, your search will work better.