Arasan 12.0 released

[jdart]

Arasan 12.0 is available from http://www.arasanchess.org. I have made some additional fixes and improvements compared to the version that played in the recent WCRCC tourney.

Changes in Arasan 12.0 (Aug. 2010):
1) Eval code is changed to split eval terms more explicitly into opening, middlegame and endgame terms and these are merged via a weighting function based on material.
2) Some eval changes to king safety, bishop pair, etc.
3) King safety cache was also used for opponent pawn proximity - but the hashing didn't account for opponent pawns. This is fixed by not storing proximity in the cache.
4) Material score adjustment added for Q vs R + Pawns, and R vs pawns.
5) Some new "trade down" bonus code is present to encourage exchanging when ahead - there was some code for this earlier but less correct.
6) Fix to futility pruning margin calculation.
7) Fixed a Winboard/xboard driver bug where some "ping" commands from ICC were being ignored.
8) Fix a rarely occurring bug in qsearch: generating a check that involves moving a pinned piece.
9) Fix bug in razoring code that could cause array overflow at very large search depths (as in endgame).
10) Fix display in GUI game list dialog & main window to include year of game

I have also updated the Arasan test suite, which now has 237 positions.

[Graham Banks]

Thanks Jon. You seem to making good progress lately.

[Werner]

Thanks Jon. You seem to making good progress lately.

Thanks too!
Just started a test!

Werner

[beachknight]

great news. will just include in my tournaments.

[Kirk]

Arasan 12.0 is available from http://www.arasanchess.org. I have made some additional fixes and improvements compared to the version that played in the recent WCRCC tourney.

Changes in Arasan 12.0 (Aug. 2010):
1) Eval code is changed to split eval terms more explicitly into opening, middlegame and endgame terms and these are merged via a weighting function based on material.
2) Some eval changes to king safety, bishop pair, etc.
3) King safety cache was also used for opponent pawn proximity - but the hashing didn't account for opponent pawns. This is fixed by not storing proximity in the cache.
4) Material score adjustment added for Q vs R + Pawns, and R vs pawns.
5) Some new "trade down" bonus code is present to encourage exchanging when ahead - there was some code for this earlier but less correct.
6) Fix to futility pruning margin calculation.
7) Fixed a Winboard/xboard driver bug where some "ping" commands from ICC were being ignored.
8) Fix a rarely occurring bug in qsearch: generating a check that involves moving a pinned piece.
9) Fix bug in razoring code that could cause array overflow at very large search depths (as in endgame).
10) Fix display in GUI game list dialog & main window to include year of game

I have also updated the Arasan test suite, which now has 237 positions.

I hope new versions of Arasan will continue for a long time!

[jdart]

couple other things I should probably mention:

Arasan now has an "easy move" detection algorithm - I used to have this in earlier versions but at some point it was removed. If a move appears clearly superior at early depths and continues to be selected, Arasan may use less time on that move.

Also - the way hash size and other memory size parameters are specified on the command switches and the arasan.rc file has changed. Hash size used to be in kilobytes but now I use "512K" for 512 kilobytes, "512M" for 512 Megabytes, etc. GUIs are not affected by this - only command line + arasan.rc.

[beachknight]

Hi,

I have run a test tournament with A12 1cpu [x64] in UCI mode on Arena.
It performed badly, scored only a half point in 18 games against 18 opponents (2600-2690).
I changed UCI to wb. The score was 6 points in 6 games (2670-2690).

I suggest therefore using Arasan 12 in wb mode.

[jdart]

Arasan has some learning features that do not work under UCI. So it may do better as a Winboard engine, but I wouldn't expect a huge difference.

--Jon

[beachknight]

Hi,

I have run a test tournament with A12 1cpu [x64] in UCI mode on Arena.
It performed badly, scored only a half point in 18 games against 18 opponents (2600-2690).
I changed UCI to wb. The score was 6 points in 6 games (2670-2690).

I suggest therefore using Arasan 12 in wb mode.

Hi Harun,

I have played 120 games with Arasan 12 under the CB GUI. It performed around 50% against suitable opposition.

Your results may be the result of a statistical anomaly. More probably, it has something to do with Arena. Take a look at the PGN, are there not many time losses? Arena is in the habit of disliking some UCI engines...

Also, make sure there is nothing in the command line (e.g. xboard).

All in all, make some survey, you may find an error somewhere.

For UCI games:

http://tech.groups.yahoo.com/group/Ches ... ssage/5649

[beachknight]

Arasan has some learning features that do not work under UCI. So it may do better as a Winboard engine, but I wouldn't expect a huge difference.

--Jon

400+ points difference between two modes. I call it huge.