Eigenmann Endgame Test: Deep Sjeng, Crafty, Nimzo a.o.

[Walter Eigenmann]

There are new results:

-Deep Sjeng 2.5
-some engines without TBs (no Nalimov-, Shredder- or Bit-Bases)
-some older engines

Regards: Walter

Code: Select all

E N D G A M E - R A N K I N G  *  April 2007/4


    Program                       Solutions    Time

01. Shredder 10                   68/100      00:45:17
02. Rybka 2.2                     64/100      00:50:48
03. Fritz 10                      63/100      00:49:47
04. Shredder 10(ohneBases)        62/100      00:52:00
05. Rybka WinFinder 2.2           61/100      00:48:54
06. Fritz 10(OhneBases)           61/100      00:52:33
07. Rybka 2.2(OhneBases)          61/100      00:53:11
08. Deep Frenzee 3.0              61/100      00:54:04
09. Deep Frenzee 3.0(OhneBases)   61/100      00:55:26
10. Shredder 9.1                  59/100      00:56:20
11. Rybka 1.01 B13d               59/100      00:56:31
12. Hiarcs 11.1                   59/100      00:58.50
13. The Baron 2.12                58/100      00:59:48
14. Hiarcs 11.1(OhneBases)        57/100      01:00:38
15. Chess Tiger 2007              54/100      00:58:55
16. Toga II 1.3x4                 54/100      01:01:58
17. Spike 1.2 Turin               53/100      00:59:42
18. SlowChess Blitz WV2.1         53/100      01:02:32
19. Delfi 5.1                     51/100      01:05:59
20. Ktulu 8                       50/100      01:02:55
21. Pharaon 3.5.1                 50/100      01:03:32
22. Yace 0.99.87                  50/100      01:04:44
23. LoopMP 12.32                  50/100      01:06:02
24. Naum 2.0                      50/100      01:06:08
25. Loop 10.32f                   49/100      01:06:25
26. Hiarcs 10                     49/100      01:08:33
27. Ruffian 2.1.0                 48/100      01:05:41
28. Toga II 1.2.1a                48/100      01:06:05
29. Fruit 2.2.1                   48/100      01:06:29
30. Crafty 20.14                  47/100      01:05:26
31. Alaric 704                    46/100      01:06:14
32. Glaurung 1.2.1                46/100      01:09:19
33. Chess Tiger 15.0              46/100      01:10:23
34. Junior 10.1                   45/100      01:06:14
35. Alaric 0702b                  45/100      01:06:54
36. Gandalf 6.0                   44/100      01:06:00
37. Deep Sjeng 2.5                44/100      01:10:49
38. Scorpio 1.91                  43/100      01:07:35
39. Patzer 3.8                    42/100      01:06:50
40. WildCat 7.0                   40/100      01:13:10
41. Anaconda 2.0.1                40/100      01:16:26
42. Junior 10                     39/100      01:12:44
43. Fritz 5.32                    39/100      01:13:11
44. Aristarch 4.50                39/100      01:13:41
45. Nimzo 8                       38/100      01:13:14
46. Pepito 1.59                   38/100      01:15:16
47. Chessmaster 9000              37/100      01:13:21
48. Chessmaster 10000             37/100      01:14:19
49. Jonny 2.83                    37/100      01:15:06
50. Nimzo 7.32                    36/100      01:14:06
51. Movei 0.08.402                35/100      01:16:22
52. Amyan 1.597                   35/100      01:16:58
53. Quark 2.35                    35/100      01:19:32
54. LGoliath Evolution            34/100      01:14:21
55. Arasan 9.5                    34/100      01:16:40
56. Gaia 3.5                      34/100      01:17:44
57. Tao 5.6                       31/100      01:19:42
58. Sjeng 12.13                   28/100      01:18:33
59. Crafty 18.01                  28/100      01:20:44
60. AnMon 5.60                    28/100      01:22:54
61. DanaSah 3.03                  27/100      01:21:13
62. SOS 5                         27/100      01:21:49
63. GreKo 5.0                     24/100      01:25:15
64. Ufim 8.02                     23/100      01:24:18
65. Hamsters 0.2                  22/100      01:22:53
66. Chispa 4.0.3                  22/100      01:25:44
67. Homer 2.0                     20/100      01:26:44
68. Queen 3.09                    19/100      01:26:43
69. Abrok 5.0                     19/100      01:28:31
70. Smirf MS-167d                 18/100      01:27:44
71. Lime 62                       17/100      01:28:33
72. AliChess 4.08                 16/100      01:27:46
73. BikJump 1.2.1                 15/100      01:27:27
74. Homer 1.1 b3                  14/100      01:32:10

(Athlon64-3000+/128MBHash/Nalimov-,Shredder-&Bit-Bases/Fritz-&Arena-GUIs)

E-E-T:
http://www.beepworld.de/members38/eigenmann/e_e_t.htm

[Jouni]

Don't forget, that Rybka WinFinder 2.2 doesn't use any bases. And some say Rybka is bad endgame player...

Jouni

[bob]

can you tell me what hardware you are using? I have posted this previously, but again, running on my dual-core 2.0ghz core-2 laptop, I get this using no endgame tables:

Code: Select all

total positions searched..........         100
number right......................          73
number wrong......................          27
percentage right..................          73
percentage wrong..................          27
total nodes searched.............. 11418366622
average search depth..............        11.4
nodes per second..................     5213175
total time........................       36:30

and this with 3-4-5 piece endgame tables:

Code: Select all

total positions searched..........         100
number right.....................          70
number wrong......................          30
percentage right..................          70
percentage wrong..................          30
total nodes searched..............  6495106143
average search depth..............        10.5
nodes per second..................     2736129
total time........................       39:33

Those are with version 21.5, which will probably do about the same as 20.x. Your numbers look way low for any reasonable version of Crafty unless your hardware is very slow...

This is the result from my old dual 2.8ghz xeon box:

Code: Select all

total positions searched..........         100 
number right......................          66 
number wrong......................          34 
percentage right..................          66  
percentage wrong..................          34
total nodes searched..............  5700572689  
average search depth..............        10.9
nodes per second..................     2288982
total time........................       41:30

[PauloSoare]

Don't forget, that Rybka WinFinder 2.2 doesn't use any bases. And some say Rybka is bad endgame player...

Jouni

What I read here and I agree, Jouni, is that all engines are not good
in endgames. In Rybka Forum , Vas talked that humans are better than
engines in endgames.

Paulo Soares

[Walter Eigenmann]

>can you tell me what hardware you are using?

See my thread start --->below

I'm using the same hardware for all engines...

Regards: Walter