Comments/Ratings for a Single Item

This year we have 8 participants; ChessV is new compared to last year, and other programs now have improved versions. Each program will play each other program 10 times, from 5 different opening arrays (Bird, Capablanca, Carrera, Embassy and an unmentionable one).

Live viewing of the games is possible at:

Live 10x8 Games

After one full round robin of Bird's Chess, the standings are:

Cross table, sorted by score percentage, Buchholz, SB 

                              Jo TS TJ Sm Fa Ch Ar Bi 
 1. Joker80 n                 ## 11 11 11 11 11 11 11  100%  14.0 ( 84.0,  84.0) 
 2. TSCP 10x8                 00 ## 0= 10 11 11 11 11   68%   9.5 ( 93.0,  42.5) 
 3. TJchess10x8 0.121         00 1= ## 01 =0 11 11 11   64%   9.0 ( 94.0,  40.3) 
 4. Smirf 1.75t               00 01 10 ## 10 11 11 11   64%   9.0 ( 94.0,  39.5) 
 5. Fairy-Max 4.8 v           00 00 =1 01 ## 1= 11 11   57%   8.0 ( 96.0,  33.5) 
 6. ChessV 0.94               00 00 00 00 0= ## 11 =1   29%   4.0 (104.0,   8.5) 
 7. ArcBishop80 1.00          00 00 00 00 00 00 ## =1   11%   1.5 (109.0,   1.5) 
 8. BigLion80 2.23x           00 00 00 00 00 =0 =0 ##    7%   1.0 (110.0,   2.8)

Rank Name                Elo    +    - games score oppo. draws 
   1 Joker80 n          2435  125   96    70   92%  1938    1% 
   2 TJchess10x8 0.121  2172   81   76    70   70%  1975    6% 
   3 Smirf 1.75t        2156   81   77    70   68%  1978    4% 
   4 TSCP something     2047   75   75    70   54%  1993    9% 
   5 Fairy-Max 4.8 v    1990   74   74    70   49%  2001    9% 
   6 ChessV 0.94        1921   73   75    70   41%  2011   10% 
   7 ArcBishop80 1.00   1642   88  103    70   14%  2051    4% 
   8 BigLion80 2.23x    1636   88  103    70   13%  2052    6% 

Note that this list only contains WinBoard-compatible engines, and engines for which a WinBoard adapter exists. There also exist non-compatible engines (e.g. Zillions). Except for Smirf, which has reverted to the status of a private engine, all mentioned engines can be downloaded for free from the internet.

After several years with no new developments in terms of engines, there was a flurry of activity last year amongst engine programmers. Five new engines appeared that can play Capablanca-like 10x8 variants, some of them very strong: 

Bihasa    by Ferdinand Mosca
Nebiyu    by Daniel Shawul
Sjaak     by Evert Glebbeek
Heretic   by Martin Sedlak
Spartacus by me

In addition the existing engines TJchess10x8 and SMIRF were improved.

As a result there are now 13 engines that can play under WinBoard, which makde it high time to conduct another automated tourney. So I am currently running 'Battle of the Goths 2012'.

For those interested in 10x8 Chess, the games can be watched live at http://80.100.28.169/gothic/chess.html .

In spite of what I see as the drawbacks of this variant (unprotected pawn for each side in setup, rectangular board [though allowing smothered and back rank mates still], bishops clearly stronger than knights, the fact the chancellors might be developed symmetrically and traded in short order sometimes), this was a good try historically to cut down on draws and opening theory.

On this particular variant's board dimensions of 10x8, as compared to 8x8, IMHO the archbishops would seem to come closer in value to chancellors (though not queens), though I personally have lingering doubts about archbishops being quite as good by comparison on 8x8 or 10x10 boards, any computer studies aside. IMHO, the bishop component of an archbishop would seem to have a number of extra potential good squares near the centre (or in range of the enemy camp) on a 10x8 board, without the rook component of a chancellor benefitting as much as often in return (unlike would be the case on a 10x10 board). On a 10x8 board the knight component of an archbishop would seem to have a number of extra potential good squares near the centre (or the enemy camp) for local scope, balancing the benefit received by the rook component of a queen on such an empty larger board than 8x8.

My tentative estimates for the piece values in this variant would be: P=1; N=3.5 approx.; B=3.75; R=5.5; A=8.25; C=10; Q=10.25 and the fighting value of the K=3.2 (though it naturally cannot be traded).

edit: Here's a 10x8 CV that uses 2 powerful and unusual pieces, besides the chess army and Berolina pawns:

https://www.chessvariants.com/play/gamma2-chess

Also, here's a 10x8 variant that uses Frogs besides the chess army:

https://www.chessvariants.com/rules/frog-chess

A link to a published preset for a circular Capablanca Chess style variant:

Circular Capa Chess

My tentative estimates for the piece values in this variant would be: P=1; N=3.625; B=3.75; R=5.5; A=8.375; C=10.125; Q=10.25 and the fighting value of the K=3.2 (though it naturally cannot be traded).

The piece values of Capablanca Chess have been measured with quite good precision. Like in orthodox Chess the Bishops do not have a single value, but have to be differentiated in 'first Bishop' (i.e. a lone one) and 'second Bishop'. This can also be expressed as a base value and a pair bonus. In orthodox Chess the base values of B and N are equal (3.25), although the B-N difference is slightly dependent on the number of remaining Pawns. (Strict equality is achieved for 5 Pawns). The pair bonus is then 0.5 Pawn.

On a 10x8 board the base values of B and N already differ by 0.5 Pawn, and the B-pair bonus just adds to this. This makes BB vs NNP an almost perfectly balanced trade, while in orthodox Chess the player with the remaining NNP would have a clear advantage, about as large as the Bishop side would have after a BB vsn NN trade. A Bishop profits from wide boards, probably because those enhance the chances that both its forward moves hit the enemy camp. In Cylinder Chess the Bishop gets even closer in value to the Rook (4 Pawns + bonus vs 5).

In your tentative estimates the Q-C difference is too small (it is 0.5 Pawn; CP vs Q is as large an advantage as Q vs C), and a the Q-A difference too large (not accounting for the fact that AP has the upper hand over Q, by about as much as C has over A).

Regarding the first version of Capablanca Chess, that is with the use of a 10x10 board, I recall reading somewhere that this was tried in testing games between Capablanca and a certain opponent (I forget who), and that the conclusion was that games tended to take (arguably) too long to finish. In any case, I'm wondering if anyone knows if in the 10x10 version, were pawns allowed to move differently than in normal chess, e.g. could a pawn take a triple step on their first move, besides a double step, if not breaking other rules of the game (similar to modern day Omega Chess' pawn rules)?

According to Pritchard, yes, the Pawn could move up to 3 spaces on first move.

Thanks Greg.

Here's the wiki re: Capablanca Chess that I think refered to earlier (re: including mention of 10x10 version):

https://en.wikipedia.org/wiki/Capablanca_Chess#Setup_and_rules

I'm wondering whether anyone has observed that it tends to be difficult for either side to castle fairly early on (if at all) in typical games of (10x8) Capablanca Chess (i.e. final version of this CV).

Quite a while back I invented a new form of castling (originally for use in my 12x8 Wide Chess CV invention). This form of castling became known as Fast Castling - assuming at least some people like this form of castling, maybe an experiment could be tried sometime where (10x8) Capablanca Chess is played with the use of Fast Castling rules (or at least posters could give their thoughts on if the change might be a good idea). The rules for Fast Castling are as follows:

"A king that has never moved, and is not in check, can 'leap' once a game, along the first rank, to any unattacked empty square between it and an unmoved rook, followed by said rook 'leaping' to the king's initial square so as to complete castling in one single move. It does not matter if any squares in between are occupied or under attack."