New chess variants

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

Ferdy
Posts: 4833
Joined: Sun Aug 10, 2008 3:15 pm
Location: Philippines

New chess variants

Post by Ferdy »

Got another 2 variants, Los Alamos on 6x6 board and s-chess or seirawan chess.

(1)
Liit v1.0

Code: Select all

Los Alamos chess variant
Readme
April 10, 2012


A. Introduction
Liit is an engine that can only play Los Alamos chess variant. This variant
is similar to chess except that it is played in 6x6 board, no bishops, no castle,
no ep capture and no double-step pawn pushes. Use a winboard gui and install this
engine to have a comfortable playing board. Read also the included FAQ file for
some other info.


B. Acknowledgement
Thanks to Michael for the help in optimizing the evaluation parameter
which resulted in a stronger play of the engine. This engine has existed due to
him encouraging to try this variant. Thanks also to H.G. Muller for the further
development of winboard, where this engine has been primarily tested.


C. Internals
Liit is based on Deuterium. It supports winboard protocol with some important 
commands implemented such as setboard, analyze, new, go, include/exclude, pondering and others. 
It can do multipv while at the same time able to exclude moves, useful for analysis.
Strength improvement is taken basically from mobility evaluation, and king safety has little
effect on this variant as proven from test.

(2)
Sigla v1.0

Code: Select all

s-chess or seirawan chess variant
Readme
April 14, 2012


A. Introduction
Sigla is an engine that can only play s-chess or seirawan chess variant. This is similar
to chess except that it has added piece types such as hawk (N+B) moves and elephant(R+N) moves.
Check the site here http://www.seirawanchess.com, for the rules and some interesting articles
from one of the inventors GM Yasser Seirawan the other is FM Bruce Harper. It is advised to
install this engine in winboard gui. Check here 
http://www.open-aurec.com/wbforum/viewtopic.php?f=19&t=51528. 


B. Acknowledgement
Thanks to H.G. Muller for the continuous development of winboard protocol and gui. The engine
has been developed based from this gui. Thanks to Tom Kerrigan for sharing the code of 
his program TSCP, and to CCC people for programming ideas.


C. Internals
Sigla is based on Deuterium engine. It supports winboard protocol with some important 
commands implemented such as setboard, analyze, new, go, and others. Its difference 
to deuterium is that it has no SEE, and relies on MVV/LVA for capture move ordering.
Search reductions and prunnings are very limited and untuned. Evaluation for new pieces
are only material values, overall evaluation is also untuned. This engine has one limitation,
it can not gate on the rook square when castling. From obsevation, engine will gate 
anyway as soon as it moves its minor pieces early in the opening. Materials are settable via
winboard gui options. There are also a strength level options where you can set from level Min
to level Max. Selecting Min allows the engine to be mated, even at level 0. There is also a
level called MyLevel where the engine attempts to always equalize your moves. If you blundered
it will also blunder. Level 6 to Max are stronger levels. Other options are Ponder, Resign,
Contempt and Book, can be set thru Winboard Engine->Engine #1 settings..... The engine can
also analyze and supports include/exclude commands.
Can be found in https://sites.google.com/site/deuterium ... ownloads-1
User avatar
hgm
Posts: 27794
Joined: Fri Mar 10, 2006 10:06 am
Location: Amsterdam
Full name: H G Muller

Re: New chess variants

Post by hgm »

I only got to downloading and testing Sigla now, because I have been too busy with the WinBoard release the past weeks. I do see some strange behavior in it, though. Look at the following game:

Code: Select all

[Event "Computer Chess Game"]
[Site "CHESS_LAPTOP"]
[Date "2012.04.21"]
[Round "-"]
[White "Schess32 12.0002 (UCI2WB)"]
[Black "Sigla v1.0"]
[Result "1-0"]
[TimeControl "40/60"]
[Variant "seirawan"]
[Annotator "1. +0.12   1... +0.06"]

1. Nc3 {+0.12/9} Nc6/H {+0.06/9 1.5} 2. Nf3 {+0.29/10 4} d6 {+0.33/8 1.5}
3. e4 {+0.67/9 2.0} Nf6/E {+0.99/7 1.5} 4. d4 {+0.99/8 1.1} d5
{+2.82/8 1.5} 5. e5 {+1.34/10 2.4} Ne4 {+1.01/7 1.5} 6. Bd3 {+1.34/9 1.2}
Nxe5 {+5.67/6 1.5} 7. dxe5 {+3.33/9 0.9} Nxc3 {+0.27/7 1.5} 8. bxc3
{+3.35/8 0.5} g5 {-1.13/7 1.5} 9. Nxg5 {+3.78/8 1.3} h6 {-1.05/6 1.5} 10.
Nxf7 {+3.78/8 1.0} h5 {-4.45/6 1.5} 11. Nxd8 {+10.88/7 0.5} Bg4
{-7.15/7 1.5} 12. f3 {+13.59/8 1.0} Bh3 {-7.01/4 1.5} 13. gxh3
{+16.76/9 1.0} c6 {-11.74/6 1.5} 14. Rg1/E {+16.86/9 1.0} Kxd8
{-14.23/8 1.5} 15. Rxg8 {+16.92/8 0.2} Rxg8 {-13.99/9 0.4} 16. f4
{+17.14/9 1.1} Rh8 {-14.53/9 1.5} 17. Be3 {+17.54/8 1.1} b5 {-14.33/8 1.5}
18. Qf3 {+18.01/9 1.7} e6 {-14.61/9 1.5} 19. O-O-O/H {+18.18/10 2.0} Kc8
{-14.62/8 1.5} 20. Eg3 {+18.32/11 1.7} Kb7 {-14.36/8 1.5} 21. Eg5
{+18.57/10 1.0} Ba3+ {-14.50/8 1.5} 22. Kb1 {+18.86/11 1.4} Hc7
{-14.62/8 1.5} 23. Ef7 {+19.23/11 4} Raf8 {-14.00/7 1.5} 24. Exc7+
{+19.55/10 3} Kxc7 {-14.89/9 0.4} 25. Hf2 {+19.63/10 4} Be7 {-14.46/7 1.6}
26. Bxa7 {+20.30/9 0.9} Ra8 {-15.64/8 1.6} 27. Qe3 {+25.07/10 2.2} c5
{-18.47/9 1.6} 28. Bxc5 {+25.52/8 0.4} Kb7 {-20.62/8 1.6} 29. Bxe7
{+28.62/8 0.2} Ra6 {-32.89/11 1.6} 30. Bxb5 {+1000.08/10 4} Rha8
{-30.76/8 1.6} 31. Bxa6+ {+1000.07/5 0.1} Rxa6 {-299.88/9 0.3} 32. Qc5
{+1000.06/5 0.1} Rc6 {-299.92/8 1.7} 33. Qb5+ {+1000.05/2 0.1} Kc7
{-299.92/8 1.7} 34. Bd6+ {+1000.04/1} Rxd6 {-299.94/6 0.7} 35. Hc5
{+1000.03/2 0.1} Kd8 {-299.96/5 0.1} 36. exd6 {+1000.02/1} d4
{-299.98/5 0.1} 37. Hxe6# {+1000.01/1 0.1}
{Xboard adjudication: Checkmate} 1-0
Sigla gates its Elephant and Hawk immediately, while its opponent keeps them in store, and quickly after that Sigla's score rises to a ridiculously high value without apparent reason. After that, it drops like a stone.

My suspicion is that it forgets to account for pieces in hand in its material evaluation, so that it gets ecstatic when it sees it can sacrifice material to push the gatings of the opponent over the horizon. As long as you still have plenty of opportunities to gate, pieces in hand could actually be worth a little more than when they are on the board.
Ferdy
Posts: 4833
Joined: Sun Aug 10, 2008 3:15 pm
Location: Philippines

Re: New chess variants

Post by Ferdy »

hgm wrote: ...
My suspicion is that it forgets to account for pieces in hand in its material evaluation,
...
Thanks for this info, looking at the source that is indeed true :( .