
Thanks
Moderator: Ras
BSG supports Winboard Protocol 2 engines only. It requires engines to be "talkative" enough and reply to ping commands.
Did you see any problem/bug when BSG plays Polyglot book moves?Torom wrote: ↑Wed Oct 13, 2021 5:02 pm For reference, why I think polyglot moves should be played by weight:
Cfish implementation: https://github.com/syzygy1/Cfish/blob/m ... polybook.c
If I understand the C code correctly, the first move with the highest weight is played as the best move.
lichess-bot implementation: https://github.com/ShailChoksi/lichess- ... ot.py#L372
If "best_move" is set, reader.find() is used from python-chess. And reader.find() is defined like this:
https://python-chess.readthedocs.io/en/ ... eader.find
"Finds the main entry for the given position or Zobrist hash.
The main entry is the (first) entry with the highest weight.
By default, entries with weight 0 are excluded. This is a common way to delete entries from an opening book without compacting it."
BSG auto matches all opening positions with ones from ECO list and it keeps the longest one to show. The ECO list is quite limited on depth thus it typically can’t cover longer openings (out of range). Since it is an auto process, users don’t have to do anything nor change it.
I have added it to my to-do list. Thanks.Nigel Robson wrote: ↑Tue Oct 26, 2021 1:10 pm The ERET test position suite contains 5 alternative move tests which Banksia skips, perhaps because they are not understood. For example
8/6pp/5k2/1p1r4/4R3/7P/5PP1/5K2 w - - am Ke2; id "ERET 098 - Endgame";
where the correct solution is any move other than Ke2. Would it be possible to add the am capability?
Thanks, Nigel
I can't find any ECO list and it never shows any classification. I tried many different book and pgn files.phhnguyen wrote: ↑Tue Oct 26, 2021 12:23 pmBSG auto matches all opening positions with ones from ECO list and it keeps the longest one to show. The ECO list is quite limited on depth thus it typically can’t cover longer openings (out of range). Since it is an auto process, users don’t have to do anything nor change it.
A bit ambiguous since I am not sure what you need:Wilhelm wrote: ↑Tue Oct 26, 2021 3:24 pmI can't find any ECO list and it never shows any classification. I tried many different book and pgn files.phhnguyen wrote: ↑Tue Oct 26, 2021 12:23 pmBSG auto matches all opening positions with ones from ECO list and it keeps the longest one to show. The ECO list is quite limited on depth thus it typically can’t cover longer openings (out of range). Since it is an auto process, users don’t have to do anything nor change it.
Do I need a ECO list?
I mean I always see {end of opening} in the notation and nothing else. Does it only work in human games?phhnguyen wrote: ↑Tue Oct 26, 2021 3:36 pmA bit ambiguous since I am not sure what you need:Wilhelm wrote: ↑Tue Oct 26, 2021 3:24 pmI can't find any ECO list and it never shows any classification. I tried many different book and pgn files.phhnguyen wrote: ↑Tue Oct 26, 2021 12:23 pmBSG auto matches all opening positions with ones from ECO list and it keeps the longest one to show. The ECO list is quite limited on depth thus it typically can’t cover longer openings (out of range). Since it is an auto process, users don’t have to do anything nor change it.
Do I need a ECO list?
- If you mean about BSG: no, it is auto and you can't interfere with the process of finding ECO codes for openings, you don't need to do any thing
- If you want to learn, collect, just Google it, there are many websites, listed a few:
https://en.wikipedia.org/wiki/Encyclopa ... s_Openings
https://www.chessgames.com/chessecohelp.html
Download:
https://github.com/jdart1/arasan-chess/ ... r/book/eco
A good result.phhnguyen wrote: ↑Tue Oct 26, 2021 1:49 pmI have added it to my to-do list. Thanks.Nigel Robson wrote: ↑Tue Oct 26, 2021 1:10 pm The ERET test position suite contains 5 alternative move tests which Banksia skips, perhaps because they are not understood. For example
8/6pp/5k2/1p1r4/4R3/7P/5PP1/5K2 w - - am Ke2; id "ERET 098 - Endgame";
where the correct solution is any move other than Ke2. Would it be possible to add the am capability?
Thanks, Nigel
BTW, how should we count the result if an engine hits an "am" move then change to other moves at the end? A good or a bad result?