Scid fork and Phalanx news

[stevenaaus]

If anyone missed my post in the tournament sections, my project Scid vs. PC is starting to look quite mature. It's got a heap of improvements over SCID - A decent gamelist widget and interface, and improvements to FICs, Tacgame, Help-menus and the Puzzle features come to mind.

The next thing to do is to port SCID's recent si4, undo(redesign?) and tree caching code. It's not a *huge* task, but any C gurus wanting to lend a hand would be welcome.

It also has a few Phalanx fixes, and i've recently chown-ed the sourceforge Phalanx project, if anyone's still interested in this engine.

[JohnS]

If anyone missed my post in the tournament sections, my project Scid vs. PC is starting to look quite mature. It's got a heap of improvements over SCID - A decent gamelist widget and interface, and improvements to FICs, Tacgame, Help-menus and the Puzzle features come to mind.

The next thing to do is to port SCID's recent si4, undo(redesign?) and tree caching code. It's not a *huge* task, but any C gurus wanting to lend a hand would be welcome.

It also has a few Phalanx fixes, and i've recently chown-ed the sourceforge Phalanx project, if anyone's still interested in this engine.

This is a great version of scid. Any chance of implementing millisecond time controls in computer matches for ultrafast matches?

Thanks.

[stevenaaus]

Any chance of implementing millisecond time controls in computer matches for ultrafast matches?

It seems to work fine now. Just input a decimal value (".2" for eg) in the time dialog, instead of using the arrows which only selects integers.

The only issue is that the xboard command is "st SECONDS". Testing phalanx, it says: "ENGINE 2 says: Error: "st:" expected integer , but got .2", so it looks like your mileage may vary. UCI time slice is given in milliseconds, so .2 becomes "go movetime 200.0" which works fine with Robo,Stock and Toga.

[Kirill Kryukov]

It also has a few Phalanx fixes, and i've recently chown-ed the sourceforge Phalanx project, if anyone's still interested in this engine.

Great to hear that Phalanx is being worked on again! Good luck, and looking forward to future releases!

Best,
Kirill

[stevenaaus]

Ha ha, it was only a two-liner to implement millisecond time control for Phalanx. I'd already fixed it's illegal castling.
You can get the latest code here

Code: Select all

svn co https://phalanx.svn.sourceforge.net/svnroot/phalanx

Cygwin is needed for windows compilation.

Things to-do : en-passant needs a bug-fix, and it's *very* weak at detecting rook mating attacks on the first and eighth rank. But we'd need a *real* chess programmer to do this.

[JohnS]

Any chance of implementing millisecond time controls in computer matches for ultrafast matches?

It seems to work fine now. Just input a decimal value (".2" for eg) in the time dialog, instead of using the arrows which only selects integers.

Great - thanks for the tip. Is it possible to specify an opening book or a set of opening positions in a pgn file for engine matches?

Thanks.

[stevenaaus]

Is it possible to specify an opening book or a set of opening positions in a pgn file for engine matches?

Sorry, but I'm a little ignorant about opening books in general. AFAIK (and can remember since i wrote this code a few months ago), the xboard engines will use their own book (if the correct directory is configured in the Tools->AnalysisEngines widget, of course) while the UCI engines are bookless at the moment. Crafty's end game book is explicitly disabled (as it breaks the analysis widget) but for the purposes of playing a tournament, it ~should~ be fine to re-enable by removing this code snippet: "egtb off".

[shiv]

Is it possible to specify an opening book or a set of opening positions in a pgn file for engine matches?

Sorry, but I'm a little ignorant about opening books in general. AFAIK (and can remember since i wrote this code a few months ago), the xboard engines will use their own book (if the correct directory is configured in the Tools->AnalysisEngines widget, of course) while the UCI engines are bookless at the moment. Crafty's end game book is explicitly disabled (as it breaks the analysis widget) but for the purposes of playing a tournament, it ~should~ be fine to re-enable by removing this code snippet: "egtb off".

Scid does support polyglot books, it would be great if those can be used during the matches. I believe they are already being used for engine annotation and the serious game features in scid.

[stevenaaus]

Scid does support polyglot books, it would be great if those can be used during the matches.

I'll certainly look at in the next couple of weeks, but i'm a little confused. Is the idea to select different books for each competitor , or have them use the same book (which would involve fewer new widgets). And wouldn't it be ok to have the matches start at a certain position (which would be easy to code too) and seems to me to suffice.

The latest release can always be got from

Code: Select all

svn co https://scidvspc.svn.sourceforge.net/svnroot/scidvspc

Since 4.1 i've added some extra code to catch malformed "feature" lines which may allow a few more engines to work (specifically, Zct).

[JohnS]

Scid does support polyglot books, it would be great if those can be used during the matches.

I'll certainly look at in the next couple of weeks, but i'm a little confused. Is the idea to select different books for each competitor , or have them use the same book (which would involve fewer new widgets). And wouldn't it be ok to have the matches start at a certain position (which would be easy to code too) and seems to me to suffice.

I would be happy to use positions such as Nunn.pgn and Silver.pgn as opening starting positions for engine matches.