Suprachess?

[Dann Corbit]

Anyone heard of this thing:
http://suprachesssystem.com/

It looks to have been compiled with Mingw

[Andres Valverde]

Seems a new UCI engine (The first Portuguese one, author claims). It doesn't show PV and is very weak (I can mate it in few moves) . The author promises new versions soon, let's see the progress...

[JuLieN]

Seems a new UCI engine (The first Portuguese one, author claims). It doesn't show PV and is very weak (I can mate it in few moves) . The author promises new versions soon, let's see the progress...

According to its author, the engine seems to use very exotic algorithms of his invention, not having anything to do with classic minimax.... hence the absence of PV?

[UncombedCoconut]

Seems a new UCI engine (The first Portuguese one, author claims). It doesn't show PV and is very weak (I can mate it in few moves) . The author promises new versions soon, let's see the progress...

According to its author, the engine seems to use very exotic algorithms of his invention, not having anything to do with classic minimax.... hence the absence of PV?

The claim is believable. My experience, FWIW: The commands the engine understands are "uci", "isready", and "position startpos moves ...". After the latter, it replies instantly with a bestmove. It seems to freeze if you set a position that's not a legal response to its last move. (If you mate it, this resets.)

With these limitations, I'm amazed that anyone got it to work in a GUI. Nevertheless, I'm curious enough to see whether it can beat a program like shatranj.py or N.E.G. (or my silly BASH engine, if I had the patience).

[SchachProfi]

Still better than another XYZ-tweak, isn't it? Computer-chess needs fresh ideas, and any approach is warmly welcome...

[UncombedCoconut]

My last post was more negative than I intended. It's always fun to see an amateur developing a board-game-playing program!
Some of the program's quirks are amusing. For instance, send it "position startpos" followed by EOF. It will try to develop each white chessman, one by one, remembering that each piece has left its original square but forgetting where it's gone.
Given black against shatranj, it held without trouble. Given white, it got a massive advantage by knowing to develop pieces. Sadly, it dropped a knight when playing this monster with 5-ply-deep tactical vision (err, after losing the advantage white had without either program's knowledge)! But of course the natural result was a threefold rep. (I mouse-slipped early, but the programs are deterministic and ignorant of GHI.)

[Event "Local Event"]
[Site "Local Site"]
[Round "1"]
[Date "2011.01.28"]
[White "Shatranj 1.17 "]
[Black "SUPRAengine.1.0="]
[Result "1/2-1/2"]

1. g3 e5 2. Nf3 e4 3. Ng1 Qe7 4. Bg2 Qe6 5. Nh3 Qc6 6. Rg1 Qe6 7. Rh1 Qc6 8.
Rg1 Qe6 9. Rh1 Qc6 10. Rg1 1/2-1/2

[Event "Local Event"]
[Site "Local Site"]
[Round "1"]
[Date "2011.01.28"]
[White "SUPRAengine.1.0="]
[Black "Shatranj 1.17 "]
[Result "*"]

1. e4 Nf6 2. e5 Ng8 3. Qf3 Nc6 4. Qf5 Nd4 5. Qd3 Nc6 6. Qd5 Nb4 7. Qb3 Nc6 8.
Nf3 Na5 9. Qc3 b6 10. b4 Nc6 11. b5 Na5 12. Bd3 Bb7 13. O-O Bd5 14. Ba3 Nb7 15.
Qd4 c6 16. bxc6 dxc6 17. Ba6 Nc5 18. Bxc5 bxc5 19. Qxc5 Qb6 20. Qxb6 axb6 21.
Bb7 Ra7 22. Bc8 Rxa2 23. Rxa2 Bxa2 24. Nc3 Bc4 25. d3 Be6 26. Bxe6 fxe6 27. Nd4
Kd7 28. Rb1 c5 29. Nf3 Kc6 30. Ng5 Nh6 31. Nxe6 Ng4 32. Nxf8 Rxf8 33. Ne4 Nxe5
34. Ng5 h6 35. Ne6 Rf5 36. Nxg7 Rg5 37. Ne6 Rg4 38. Nd8+ Kd7 39. Nb7 Kc7 40.
Kf1 Kxb7 41. h3 Rh4 42. Re1 Rh5 43. f4 Nc6 44. c4 Rh4 45. Re4 Rh5 46. Re6 Rh4
47. Re4 Rh5 48. Re6 *