About 90% of Chess Engines Can Play Shuffle Chess

[Spock]

Unfortunately Stockfish crashes after a castling, while Glaurung can play there!?

FRC support in Stockfish is broken. They have identified the problem, but there is no sign of when they are releasing a new version.

[Alexander Schmidt]

Why should I do tests from illegal positions?

Also if an engine doesnt crash or hang, you never know how it internally reacts. So the results are useless for serious testing.

If you enjoy watching this games, do it

[Alexander Schmidt]

The castling rules in chess960 are slightly strange.

They are easy if u understood it once

The castling ends always on the fields like in normal chess. You cannot castle through check, the fields between king and rook and the castling fields must be empty.

[Spock]

the fields between king and rook and the castling fields must be empty.

I'm not sure I understand this - in FRC, the very first move can be a castle

See white's castle here

[Event "CCRL 404FRC"]
[Site "CCRL"]
[Date "2006.08.22"]
[Round "1.1.95"]
[White "Spike 1.2 Turin"]
[Black "Glaurung 1.2.1"]
[Result "1/2-1/2"]
[Opening "BQNRNKRB"]
[PlyCount "139"]
[SetUp "1"]
[FEN "bqnrnkrb/pppppppp/8/8/8/8/PPPPPPPP/BQNRNKRB w GDgd -"]
[WhiteElo "2818"]
[BlackElo "2740"]

1. O-O {+0.15/13 10s} b5 {+0.09/14 17s} 2. b4 {+0.16/12 10s} g6 {+0.10/15 24s}
3. Bxh8 {+0.45/13 7s} Rxh8 {+0.06/15 16s} 4. g3 {+0.38/13 12s} Bxh1 {+0.04/15
16s} 5. Kxh1 {+0.54/14 8s} Nf6 {+0.03/14 17s} 6. Nf3 {+0.42/14 12s} h6
{+0.00/14 19s} 7. Nb3 {+0.54/13 8s} Nb6 {+0.00/14 12s} 8. Qa1 {+0.61/13 11s}
Kg7 {-0.03/13 15s} 9. d3 {+0.72/13 8s} d6 {+0.09/13 18s} 10. Nbd4 {+0.95/13 8s}
Qb7 {+0.14/13 9s} 11. Nxb5 {+1.18/13 7s} Nbd7 {+0.32/12 10s} 12. c4 {+0.87/15
17s} Ne5 {+0.21/13 5s} 13. Nbd4 {+0.90/15 19s} Nxf3 {+0.21/14 11s} 14. Nxf3
{+0.75/15 7s} Qxb4 {+0.29/15 12s} 15. Rb1 {+0.80/16 11s} Qa4 {+0.34/15 6s} 16.
Rb7 {+0.90/14 7s} Rd7 {+0.34/14 6s} 17. Rfb1 {+0.92/14 7s} e5 {+0.26/14 22s}
18. Qb2 {+0.90/13 7s} Re8 {+0.21/12 5s} 19. e4 {+0.82/13 9s} c6 {+0.18/13 4s}
20. Rb3 {+0.68/13 8s} Qa5 {+0.20/13 5s} 21. Kg2 {+0.60/12 7s} Rde7 {+0.10/13
6s} 22. Ra3 {+0.45/13 7s} Qc7 {+0.12/13 5s} 23. Qd2 {+0.46/13 6s} Nd7 {-0.07/13
11s} 24. Rab3 {+0.30/14 7s} Nc5 {-0.07/14 4s} 25. Rb4 {+0.25/14 10s} Qa5
{-0.17/12 6s} 26. Kf1 {+0.07/13 16s} Qa3 {-0.37/12 5s} 27. Ne1 {+0.19/13 9s}
Rf8 {-0.34/12 4s} 28. f3 {+0.13/13 9s} Rd7 {-0.26/11 7s} 29. d4 {+0.40/12 9s}
exd4 {-0.34/12 4s} 30. Qxd4+ {+0.31/13 7s} f6 {-0.29/11 2s} 31. Qd2 {+0.34/12
6s} Rfd8 {-0.50/11 5s} 32. Qf2 {+0.16/11 7s} Re8 {-0.42/11 7s} 33. g4 {+0.10/11
5s} Qc3 {-0.65/11 7s} 34. Qc2 {-0.18/12 7s} Qe5 {-0.54/11 2s} 35. R4b2
{-0.21/12 5s} Ne6 {-0.67/11 6s} 36. Rb7 {+0.33/12 6s} Rde7 {-0.40/11 2s} 37.
Nd3 {+0.36/12 5s} Qd4 {-0.59/12 2s} 38. R1b3 {+0.06/11 4s} a5 {-0.45/10 2s} 39.
h4 {+0.22/11 5s} Kf7 {-0.40/9 1s} 40. Nc1 {+0.05/10 5s} Nc5 {-0.73/10 1s} 41.
Rxe7+ {-0.16/13 12s} Rxe7 {-0.54/14 18s} 42. Rc3 {-0.24/13 8s} Nxe4 {-0.64/13
11s} 43. fxe4 {+0.55/13 10s} Rxe4 {-0.65/14 16s} 44. h5 {+0.47/12 8s} Rf4+
{-0.62/13 15s} 45. Ke1 {+0.71/12 12s} Qg1+ {-0.45/13 15s} 46. Kd2 {+0.60/13
29s} Rd4+ {-0.48/12 9s} 47. Rd3 {+0.60/13 24s} gxh5 {-0.48/13 13s} 48. Ne2
{+0.92/13 19s} Rxd3+ {-0.60/11 2s} 49. Qxd3 {+0.90/13 7s} Qxg4 {-0.59/12 4s}
50. Qxd6 {+0.79/12 12s} Qxc4 {-0.59/11 2s} 51. Qc7+ {+0.80/12 10s} Kg6
{+0.00/13 14s} 52. Qxa5 {+0.98/12 10s} h4 {+0.00/13 15s} 53. Ke3 {+1.21/12 7s}
c5 {+0.00/13 11s} 54. Qa8 {+1.37/10 7s} Kf7 {+0.04/12 10s} 55. a4 {+1.61/10 9s}
Qb3+ {+0.18/12 10s} 56. Kf2 {+1.66/10 11s} h3 {+0.26/11 21s} 57. Qe4 {+1.70/10
6s} c4 {+0.28/10 5s} 58. a5 {+1.71/10 12s} Qa2 {+0.28/11 9s} 59. Qd5+ {+1.72/11
7s} Ke7 {+0.28/11 5s} 60. Qb7+ {+1.72/12 9s} Kd6 {+0.23/12 7s} 61. Qb8+
{+1.52/11 5s} Kd7 {+0.23/12 14s} 62. Qb5+ {+1.53/12 8s} Kc7 {+0.23/12 6s} 63.
Qb6+ {+1.68/11 8s} Kd7 {+0.21/11 4s} 64. Qxf6 {+1.68/11 6s} Qxa5 {+0.17/11 6s}
65. Qd4+ {+1.68/11 8s} Kc6 {+0.00/10 6s} 66. Qxc4+ {+1.70/10 11s} Qc5+
{+0.00/13 6s} 67. Nd4+ {+0.23/13 15s} Kd6 {+0.00/14 6s} 68. Qe6+ {+0.00/12 12s}
Kc7 {+0.00/5 0s} 69. Qe4 {+0.00/12 4s} Kd6 {+0.00/13 7s} 70. Qe6+ {+0.00/12 4s}
1/2-1/2

[IWB]

Hello

Yes, all is ok. AFTER the castling King and Rook are on the normal positions like in normal chess. For the castling itself we have the same rules as in normal chess (so no move over a field in check, no jump over third pieces ...) As I wrote there is only one exception, the king can pass the original field of the castling rook.

EG: [d]rnkbbrqn/pppppppp/8/8/8/8/PPPPPPPP/RNKBBRQN w FAfa -

Here the 0-0 would end with the King in the g-file which means it passes the original rook position. (Of course all pieces in between have to be off the fields - as in normal chess.

Bye
Ingo

[IWB]

Hello

Shuffle chess is almost similar to FRC, the only difference is the castling. ...

For me the biggest difference is that Shuffle chess has complete shuffling of all pieces (no pawns). This means the biggest difference to FRC 960 and normal chess is that it allows 2 equaly colored bishops. Do you want that?

Bye
Ingo

[swami]

Hello

Shuffle chess is almost similar to FRC, the only difference is the castling. ...

For me the biggest difference is that Shuffle chess has complete shuffling of all pieces (no pawns). This means the biggest difference to FRC 960 and normal chess is that it allows 2 equaly colored bishops. Do you want that?

Bye
Ingo

I'd prefer opposite colored Bishops but shuffled positions like in FRC. That's what Arena generates, no?

[IWB]

Hi there

I'd prefer opposite colored Bishops but shuffled positions like in FRC. That's what Arena generates, no?

I dont know what Arena does, but if it does it is 960! All the engines have to do is implemt castling.

Shredder Classic and Chessbase are creating perfect 960 positions and games.

Bye
Ingo

[Spock]

Some more info here

http://www.chessbox.de/Compu/fullchess5_e.html

[hgm]

There are several ways of shuffling. Color-bound pieces like Bishops are almost always kept on differently colored squares. In games with normal castling, Kings and Rooks are not shuffled (variant wildcastle). In games with FRC-style castling, they are shuffled, but K must remain between R (FRC, CRC). In games without castling, the King and Rooks are shuffled like any other.

Then ICC also supports a variant where the white and black shufflings are independent.

Finally, ICC has a variant that is not really shuffling, but where the pieces are randomly drawn (but equal for both sides). So you can start with 3Q, 2B, 1R, 1N. In that case the Bishops (if you have more than one) can be on like colors.

WinBoard does not support the latter two variants in local play.