First of all I decided to leave Negamax and go with Principal Variation Search (PVS). Also decided to implement separate methods for root nodes, PV nodes and null window searches. On previous experiments PVS was performing worse than Negamax, but I discovered the reason: the aspiration window has some implementation issues with PVS: when the search for a move fails low at the root node the move must be researched enlarging the window.

I was very stranged of why Futiliy Pruning was not working for me, but finally discovered the reason: a simple sign change after evaluation was the reason! Also implemented to store the evaluation values on the Transposition Table (TT).

The next step: why Carballo searched much less depths than other engines, it was due to quiescence search. I was generating checks for the first 4 PLY’s of quiescence, but some other engines not, so this was the reason. I decide to generate only checks on the first PLY of quiescence and only for PV nodes. Also modified a bit the move generation to optimize for quiescence.

During this time also found many interesting bugs, I was storing on the TT the bound and not the score when failing high/low, also on PV nodes is better to use only the TT for ordering and not to return scores from it, this helps avoiding draws. Also found a serious bugs involving time management (was taking as reference opponent’s time) and with contempt factor on IID searches corrupting TT entries. Finally added a Pawn Push Extension and removed the Recapture Extension and now some extensions now depend of the node type.

Running some test tournaments, I hope to get some good results soon and add the improved engine to my Mobialia Chess.

Jim Ablett has published a packaged executable for my Carballo chess engine. He also made a pretty logo for it, shown here. Check his page to download executables for Carballo and a lot of other chess engines: Jim Ablett’s Winboard Chess Page. Thanks Jim for your great work!

Also has many improvements on the interface and gameplay, which makes Mobialia Chess 2.0 the best chess app for Android.

Try it! The LITE version is free, and the complete version is avaiable for only 1.99 EUR at the Android Market.

Check all the Mobialia Chess for Android features at http://www.mobialia.com/chess

GWT is provided from Google to develop applications in Javascript programming in Java. It conterts the Java code to Javascript, and the application is run on the browser, without any additional plugin.

This interface is quite better than the sample Applet from Carballo, it has move history, legal move hightlighting, and SVN graphics. It can be used from:

http://www.vectomatic.org/gwt/libgwtsvg-chess/libgwtsvg-chess.html

He also wrote an article explaining the difficulties of his project. Thanks Lukas for your work.

• Static Exchange Evaluator (SEE): the biggest change,  affects move ordering introducing more move generation phases, also affects LMR, etc.
• New “experimental” evaluation function, with King Safety, X-Ray attacks and improved mobility
• Recapture extension, also changed a lot the extension mechanics adding fractional extensions
• Quiescense search now only generate checks the first 4 PLYs and follows only good captures
• UCI Options: can configure Hash Size, Book, LRM, IID, extensions, the evaluator to use, etc.
• Bugs with draw detection by the 50 move rule (I was doing at 50 PLY), and lots of other small bugs solved

Finally the Negascout code did not improve results at tournaments, so all the Negascout code is commented in the SearchEngine. I’ll continue to investigate.

With this release I include an package to run the program from chess GUIs like Arena, so now it can be included in “official” tournaments.

ELO at tests had a big increase from last version. Now scores over 2330 ELO points at BT2630 test (from the 2100 of Carballo 0.2). Also the positions solved of the “win at chess” test are now 275 of 300 (from the previous 237/300), well, also helped that the EPD parser had a bug and didn’t work well with positions with more than one solution. In LCT II also scores 2300 ELO. All the tests results are in the SVN, at the folder “testresults”.

But at tournament (playing against other java chess engines), the improvement is not so much. Now I think is over BrembroCE, but remains behind Jonathan Pettersson’s Mediocre, Alf and Frank Walter chess engines. I will center the next testing in tournament play.

Play Against Carballo Chess Engine

Play Against Carballo Chess Engine

It is a work-in progress, but a good framework to test some ideas in chess programming and good enought to play some interesting games against it (well, actually I am not able to win it, but I am a very poor chess player).

Now I’m mounting my kitchen from Ikea, so I haven’t time to improve Carballo, but in a near future I’ll continue working on it.