Hippocampe 0.4.1 released

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

Moderator: Ras

Jérémy Pages
Posts: 11
Joined: Sun Aug 17, 2008 2:28 am
Location: Mulhouse (Alsace, France)

Hippocampe 0.4.1 released

Post by Jérémy Pages »

The engine is available on my FTP server : ftp://chessgeek.dyndns.org

Here are two files :
- hippocampe-0.4.1.tar.gz (for GNU/Linux and other Unices)
- hippocampe-0.4.1-win.tar.gz (for Windows).

The 0.4 version is also available, but a bit buggy.

The program is free software, with a GPLv3 license. It works with the Xboard protocol.
Its strength is currently low, but learning and optimization algorithms will be implemented in future versions.
Philippe

Re: Hippocampe 0.4.1 released

Post by Philippe »

Can't connect
Dann Corbit
Posts: 12792
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: Hippocampe 0.4.1 released

Post by Dann Corbit »

I did a windows build using MS VC++. I may have goofed something up, but Hippocampe plays rather strange for me. He does not understand checkmate, for instance. He also thrusts his queen forward at the earliest possible chance. I guess he is named for the brain parts.

Code: Select all

[Event "Hip"]
[Site "T3400"]
[Date "2008.10.02"]
[Round "1"]
[White "Hippocampe"]
[Black "Hippocampe"]
[Result "1-0"]
[BlackElo "2200"]
[ECO "A18"]
[Opening "English"]
[Time "16:18:02"]
[Variation "Mikenas, Flohr Variation"]
[WhiteElo "2200"]
[TimeControl "120+5"]
[Termination "time forfeit"]
[PlyCount "123"]
[WhiteType "program"]
[BlackType "program"]

1. c3 g6 2. e4 a5 3. Qg4 e6 4. a3 a4 5. Ra2 Nc6 6. Ra1 Na5 7. d3 Nf6 8. Qh4
Nb3 9. e5 Nxa1 10. exf6 Ra6 11. g4 Ra7 12. Kd1 Bd6 13. Ke1 Nb3 14. Be3 Ra8
15. c4 Be5 16. Bg5 Bxb2 17. Bg2 Bd4 18. Ne2 Ba1 19. Kd1 Bb2 20. Ng3 Ra7 21.
f4 Ra6 22. Bf1 c5 23. Ne2 Rb6 24. Ng3 Rd6 25. Bg2 h6 26. Bf1 Nd4 27. Ne4
Rb6 28. Bxh6 d5 29. Nxc5 Nc6 30. Ke2 dxc4 31. dxc4 Bxf6 32. g5 Rb2+ 33. Kf3
Be7 34. Nxb7 Bxb7 35. Kg4 Qd1+ 36. Kg3 Nd4 37. Qg4 Rb3+ 38. Nc3 Rxc3+ 39.
Bd3 Rxd3+ 40. Kf2 Qxg4 41. f5 Qg2+ 42. Ke1 Ba8 43. Bf8 Rd2 44. c5 Rh5 45.
Rf1 Rxh2 46. Rf3 Bxf3 47. Bg7 Bd5 48. Bh6 Ba2 49. fxe6 Bxe6 50. Bg7 Rc2 51.
Be5 Nb5 52. Kd1 Qxg5 53. Bg7 Nd6 54. c6 Bh3 55. Ke1 Rc4 56. Bf8 Rc5 57. Kd1
Ne4 58. Bg7 Rb5 59. Ke1 Nf6 60. Bf8 Nh7 61. Bg7 Rb3 62. Bh8 ... {1-0 Black
forfeits on time} 1-0
Here is the one that I made (builds with MS VC++):
http://cap.connx.com/chess-engines/new- ... ocampe.zip
It's a 64 bit version
Jérémy Pages
Posts: 11
Joined: Sun Aug 17, 2008 2:28 am
Location: Mulhouse (Alsace, France)

Re: Hippocampe 0.4.1 released

Post by Jérémy Pages »

I may have goofed something up, but Hippocampe plays rather strange for me. He does not understand checkmate, for instance. He also thrusts his queen forward at the earliest possible chance. I guess he is named for the brain parts.
The strange style of play comes from the simplistic eval function used (it counts only material, with the following bareme : 100 for a pawn, 300 for a knight, 325 for a bishop,
500 for a rook and 900 for a queen). The mate has a value of 32000, but it does not
allow the program to distinguish mates in 1 and mates in more moves. In consequence,
it does not play properly when a forced mate exists.

However, the loss on time is more surprising for me.

The name can have several origins : the brain part, or the fish (my engineering school's symbol) named "seahorse" in english.
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Re: Hippocampe 0.4.1 released

Post by sje »

Jérémy Pages wrote:The mate has a value of 32000, but it does not
allow the program to distinguish mates in 1 and mates in more moves.
You should fix this soon. To see why, check out the game between COKO III and Genie in the 1971 ACM Computer Chess Championship.
gerold
Posts: 10121
Joined: Thu Mar 09, 2006 12:57 am
Location: van buren,missouri

Re: Hippocampe 0.4.1 released

Post by gerold »

Philippe wrote:Can't connect
No connect here either.
plattyaj

Re: Hippocampe 0.4.1 released

Post by plattyaj »

nslookup for chessgeek.dyndns.org is 192.168.1.10 which is an internal TCP/IP address and shouldn't be on the public internet at all. That's certainly why my PC can't route to it.

Andy.
User avatar
Ovyron
Posts: 4562
Joined: Tue Jul 03, 2007 4:30 am

Re: Hippocampe 0.4.1 released

Post by Ovyron »

Jérémy Pages wrote:900 for a queen
A value of 975 for the queen would be more accurate.
User avatar
JuLieN
Posts: 2949
Joined: Mon May 05, 2008 12:16 pm
Location: Bordeaux (France)
Full name: Julien Marcel

Re: Hippocampe 0.4.1 released

Post by JuLieN »

Image

Now that we have hippocamps we're looking forward to the Cartman engine.
"The only good bug is a dead bug." (Don Dailey)
[Blog: http://tinyurl.com/predateur ] [Facebook: http://tinyurl.com/fbpredateur ] [MacEngines: http://tinyurl.com/macengines ]
Dann Corbit
Posts: 12792
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: Hippocampe 0.4.1 released

Post by Dann Corbit »

plattyaj wrote:nslookup for chessgeek.dyndns.org is 192.168.1.10 which is an internal TCP/IP address and shouldn't be on the public internet at all. That's certainly why my PC can't route to it.

Andy.
Here is the one that I made (builds with MS VC++):
http://cap.connx.com/chess-engines/new- ... ocampe.zip