Tcheran bug?

[Werner]

28. f4 +8,76 ?? and then disconnects ..

[Event "Tcheran"]
[Site "?"]
[Date "2025.11.29"]
[Round "29"]
[White "Tcheran 8.0 x64 1CPU"]
[Black "bitbit 1.6 x64 1CPU"]
[Result "0-1"]
[ECO "D15"]
[PlyCount "56"]
[EventDate "2025.??.??"]
[TimeControl "600+5"]

1. d4 {book} d5 {book} 2. c4 {book} c6 {book} 3. Nf3 {book} Nf6 {book} 4. Nc3 {
book} a6 {book} 5. c5 {book} Nbd7 {book} 6. Bf4 {0.43/20 13} Nh5 {-0.15/21 20}
7. Bd2 {0.54/20 16} Nhf6 {-0.18/21 11} 8. g3 {0.38/21 14} e5 {-0.20/20 8.2s} 9.
dxe5 {0.97/22 14} Ne4 {-0.27/22 13} 10. Nxe4 {1.00/23 16} dxe4 {-0.27/22 9.4s}
11. Ng5 {0.77/22 14} Nxc5 {-0.14/22 13} 12. Bg2 {0.62/22 22} Bf5 {-0.22/22 14}
13. b4 {0.73/22 25} Ne6 {-0.24/22 16} 14. Nxe6 {0.69/22 11} fxe6 {-0.19/22 13}
15. O-O {0.49/21 22} Be7 {-0.19/22 13} 16. Rc1 {0.66/20 13} Qc7 {-0.20/21 20}
17. Qb3 {0.85/19 10} O-O {-0.24/21 10} 18. Bf4 {1.02/18 10} Rad8 {-0.26/20 8.3s
} 19. Rc4 {0.63/22 35} g5 {-0.19/21 11} 20. Be3 {0.53/22 11} Qxe5 {-0.20/22 24}
21. Qc2 {0.45/21 34} h5 {-0.24/21 8.2s} 22. Qc1 {0.54/19 15} Bg6 {-0.17/21 18}
23. a4 {0.48/20 24} h4 {-0.14/21 11} 24. Bb6 {0.72/22 13} Rd5 {-0.11/21 15} 25.
Qe3 {0.48/22 16} g4 {0.28/20 9.4s} 26. Qh6 {0.00/22 20} Kf7 {0.56/19 8.0s} 27.
Qe3 {0.00/22 13} Bg5 {0.92/20 7.4s} 28. f4 {8.76/20 8.5s} gxf3 {abandoned +2.
83/20 7.6s, White disconnects} 0-1

[jgilchrist]

Hi Werner, thanks for reporting. Is there a log file next to the engine binary? It should hopefully have some details about what went wrong.

[jgilchrist]

Actually, no need. Tcheran apparently doesn't consider gxf3 a legal move - pretty incredible, I don't think this is a new bug as I haven't changed anything about move generation in many releases.

Thanks again for reporting - I'll see what I can do about fixing this

[jgilchrist]

Confirmed that this has been an issue since essentially v1.0! Tcheran uses legal movegen and thinks that black will be putting itself in check by taking en-passant. Amazed that this hasn't presented itself before now.

[Ajedrecista]

Hello:

I enclose the game into PGN tags to get a better view of the game:

[pgn][Event "Tcheran"]
[Site "?"]
[Date "2025.11.29"]
[Round "29"]
[White "Tcheran 8.0 x64 1CPU"]
[Black "bitbit 1.6 x64 1CPU"]
[Result "0-1"]
[ECO "D15"]
[PlyCount "56"]
[EventDate "2025.??.??"]
[TimeControl "600+5"]

1. d4 {book} d5 {book} 2. c4 {book} c6 {book} 3. Nf3 {book} Nf6 {book} 4. Nc3 {
book} a6 {book} 5. c5 {book} Nbd7 {book} 6. Bf4 {0.43/20 13} Nh5 {-0.15/21 20}
7. Bd2 {0.54/20 16} Nhf6 {-0.18/21 11} 8. g3 {0.38/21 14} e5 {-0.20/20 8.2s} 9.
dxe5 {0.97/22 14} Ne4 {-0.27/22 13} 10. Nxe4 {1.00/23 16} dxe4 {-0.27/22 9.4s}
11. Ng5 {0.77/22 14} Nxc5 {-0.14/22 13} 12. Bg2 {0.62/22 22} Bf5 {-0.22/22 14}
13. b4 {0.73/22 25} Ne6 {-0.24/22 16} 14. Nxe6 {0.69/22 11} fxe6 {-0.19/22 13}
15. O-O {0.49/21 22} Be7 {-0.19/22 13} 16. Rc1 {0.66/20 13} Qc7 {-0.20/21 20}
17. Qb3 {0.85/19 10} O-O {-0.24/21 10} 18. Bf4 {1.02/18 10} Rad8 {-0.26/20 8.3s
} 19. Rc4 {0.63/22 35} g5 {-0.19/21 11} 20. Be3 {0.53/22 11} Qxe5 {-0.20/22 24}
21. Qc2 {0.45/21 34} h5 {-0.24/21 8.2s} 22. Qc1 {0.54/19 15} Bg6 {-0.17/21 18}
23. a4 {0.48/20 24} h4 {-0.14/21 11} 24. Bb6 {0.72/22 13} Rd5 {-0.11/21 15} 25.
Qe3 {0.48/22 16} g4 {0.28/20 9.4s} 26. Qh6 {0.00/22 20} Kf7 {0.56/19 8.0s} 27.
Qe3 {0.00/22 13} Bg5 {0.92/20 7.4s} 28. f4 {8.76/20 8.5s} gxf3 {abandoned +2.
83/20 7.6s, White disconnects} 0-1[/pgn]

Then, the position before 28.- f4 was:

[d]5r2/1p3k2/pBp1p1b1/3rq1b1/PPR1p1pp/4Q1P1/4PPBP/5RK1 w - - 4 28

Some perft results computed with gperft 1.1 perft counter are:

Code: Select all

5r2/1p3k2/pBp1p1b1/3rq1b1/PPR1p1pp/4Q1P1/4PPBP/5RK1 w - - 4 28

perft(1) =              41
perft(2) =           1,894
perft(3) =          76,404
perft(4) =       3,550,068
perft(5) =     142,117,684
perft(6) =   6,526,708,365
perft(7) = 260,422,448,192

Regards from Spain.

Ajedrecista.

[jgilchrist]

I enclose the game into PGN tags to get a better view of the game:

Thanks! The bug is solved, and will be released with the next version of Tcheran: https://github.com/tcheran-chess/tchera ... 4964837f27