Fruit 051103 - Strong Setting by Claudio Fabiano!

[Edward German]

Fruit 051103 - Strong Setting by Claudio Fabiano!

Read the original message here:

http://f50.parsimony.net/forum202220/messages/26558.htm

e.g. written by Claudio Fabiano:

"Fruit 051103 (history depth1=2) seems to be very strong."

----------

In my Folder and in the *.uci it looks so:

[ENGINE]
Name=Fruit 05/11/03e2
Author=Fabien Letouzey
Filename=C:\Programme\ChessBase\Engines\Fruit-051103\Fruit-051103.exe
[OPTIONS]
History Depth 1=2 (default=1)

For my tests I named the setting "Fruit 05/11/03e2".

First I test it with my testpositions, and it looks very good! Then I played an short Blitz-Match against Rybka 1.2f UCI (Nunn1 with 10 variations = 20 games) 6m + 3s.

GUI= Fritz GUI8
PB=OFF
TB=ON (3+4+5 pieces)
Hash=16 MB, 16 MB
PC= My P3 600 MHz

Result after 20 Nunn1 games:

Code: Select all

P3 600, Blitz:6'+3"  0

1   Rybka 1.2f 32-bit   1½½½110½½1½½½½00½½½0   10.0/20  100.00
2   Fruit 05/11/03e2    0½½½001½½0½½½½11½½½1   10.0/20  100.00

Result ended draw after game 20. I think, that is not bad for the last Fabien "Fruit-Beta", or?

Example:

Round 15:

[Event "P3 600, Blitz:6' + 3""]
[Site "P3 600 MHz"]
[Date "2007.08.01"]
[Round "15"]
[White "Fruit 05/11/03e2"]
[Black "Rybka 1.2f 32-bit"]
[Result "1-0"]
[ECO "E98"]
[Annotator "0.46;0.24"]
[PlyCount "117"]
[TimeControl "360+3"]

{W=14.0 ply; 155kN/s B=9.6 ply; 22kN/s} 1. d4 Nf6 2. c4 g6 3. Nc3 Bg7 4. e4 d6
5. Be2 O-O 6. Nf3 e5 7. O-O Nc6 8. d5 Ne7 9. Ne1 Ne8 10. Be3 {
Beide letzter Buchzug} f5 {0.24/12 24} 11. f3 {0.46/12 15} Bd7 {
(Kh8) 0.23/11 23} 12. Nd3 {0.25/12 18} b6 {(f4) 0.23/11 34} 13. g4 {
(Db3) 0.40/12 30} f4 {0.13/11 16} 14. Bf2 {0.24/13 15} Nf6 {(Kh8) 0.12/11 66}
15. g5 {(Db3) 0.44/13 19} Nh5 {0.08/8 2} 16. Kh1 {0.52/12 9} c5 {0.08/11 10}
17. a3 {(Tb1) 0.41/11 12} a6 {(Kh8) 0.05/10 13} 18. b4 {0.38/11 10} Rb8 {
(Kh8) 0.08/10 10} 19. Qb3 {(Tg1) 0.49/12 14} Qc7 {(Kh8) 0.09/10 13} 20. Rab1 {
(Tfd1) 0.48/11 10} Rfc8 {(Kh8) 0.13/10 11} 21. Qc2 {(Tfd1) 0.48/11 15} Qa7 {
(Kh8) 0.13/9 12} 22. Rb3 {(Tg1) 0.76/11 20} cxb4 {(a5) 0.22/10 11} 23. Nxb4 {
(Txb4) 0.93/12 44} h6 {(Kh8) 0.19/10 12} 24. gxh6 {1.46/11 10} Bxh6 {0.29/10 3}
25. Rfb1 {(Db2) 1.43/11 9} a5 {(Le8) 0.54/9 8} 26. Nd3 {(Sba2) 2.11/11 11} Rxc4
{(Le8) 1.15/10 10} 27. Nxe5 {2.58/11 12} Rcc8 {1.43/10 7} 28. Nc4 {2.35/10 5}
Bg7 {1.33/9 8} 29. a4 {(Sxd6) 2.54/10 10} Kh7 {1.23/9 14} 30. Qa2 {
(Lxb6) 2.89/11 30} Rxc4 {(Lf6) 1.19/8 8} 31. Bxc4 {3.05/9 2} Qc7 {1.61/9 9} 32.
Bd3 {3.37/10 6} Nc8 {(Le5) 1.75/9 11} 33. Nb5 {(Tc1) 3.45/11 7} Qd8 {1.67/9 6}
34. Nd4 {(Ta3) 3.50/11 8} Qe8 {(Lxd4) 1.79/10 7} 35. Bb5 {(Ta3) 3.56/11 4} Kg8
{(Lxd4) 1.98/9 11} 36. Bxd7 {(Se6) 3.92/12 5} Qxd7 {2.13/10 3} 37. Nc6 {
(Se6) 3.89/13 8} Rb7 {(Ta8) 2.32/10 18} 38. Nxa5 {(Lxb6) 4.17/12 6} Ra7 {
(Tc7) 2.26/10 7} 39. Nc6 {4.36/11 7} Ra8 {2.27/10 13} 40. Bxb6 {4.66/11 6} Nxb6
{2.55/10 19} 41. Rxb6 {4.79/10 4} Kh7 {(Dh3) 2.73/12 5} 42. a5 {(Tb8) 5.18/11 5
} Qh3 {2.73/10 4} 43. Qg2 {5.78/12 7} Qh4 {2.73/11 3} 44. a6 {(Kg1) 5.75/12 7}
Kh6 {2.93/8 6} 45. Rd1 {(Dg4) 6.66/11 19} Nf6 {4.11/9 4} 46. a7 {
(Tg1) 7.55/11 5} Ne8 {5.02/12 4} 47. Rb8 {7.88/11 3} Nc7 {5.18/12 2} 48. e5 {
(Tg1) 7.74/11 5} Bf8 {(Txa7) 6.08/11 5} 49. exd6 {(e6) 11.28/10 6} Bxd6 {
13.61/13 13} 50. Qb2 {17.25/10 3} Qh5 {15.10/13 10} 51. Qh8+ {22.10/9 2} Kg5 {
#8/5 0} 52. Rg1+ {22.10/10 6} Kh4 {#7/4 0} 53. Rg4+ {(Txa8) #7/9 4} Kh3 {
14.12/5 0} 54. Rxa8 {#5/11 4} Bf8 {(Sxa8) #6/7 0} 55. Rxf8 {(Dxh5+) #4/34 1}
Nxd5 {#5/3 0} 56. Ne5 {(Dxh5+) #3/41 4} Ne3 {(Se7) #3/3 0} 57. Nxg6 {#3/40 3}
Ng2 {#2/3 0} 58. Qxh5+ {#2/38 3} Nh4 {#1/3 0} 59. Rxh4# {#1/38 3} 1-0

Strong played by "Fruit 05/11/03e2" (Setting by Claudio Fabiano)!

In last game 20 Fruit draws the score then to 10 - 10!

[Event "P3 600, Blitz:6' + 3""]
[Site "P3 600 MHz"]
[Date "2007.08.01"]
[Round "20"]
[White "Rybka 1.2f 32-bit"]
[Black "Fruit 05/11/03e2"]
[Result "0-1"]
[ECO "B97"]
[Annotator "0.12;0.23"]
[PlyCount "152"]
[TimeControl "360+3"]

{W=10.6 ply; 27kN/s; 2.974 TBAs B=14.3 ply; 183kN/s; 1.031 TBAs} 1. e4 c5 2.
Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Bg5 e6 7. f4 Qb6 8. Qd2 Qxb2 9. Rb1
Qa3 {Beide letzter Buchzug} 10. Be2 {0.12/10 17} Nc6 {0.23/11 11} 11. Nxc6 {
0.14/11 21} bxc6 {-0.03/10 7} 12. O-O {(Tb3) 0.12/11 16} Be7 {0.14/11 16} 13.
Bh4 {(Tb3) 0.09/11 37} d5 {(0-0) -0.24/11 19} 14. e5 {0.12/9 6} Nd7 {
(Lc5+) -0.32/12 17} 15. Rb3 {0.03/10 13} Bc5+ {(Dc5+) -0.35/12 12} 16. Kh1 {
0.03/12 19} Qa5 {-0.29/12 8} 17. Rfb1 {0.03/11 10} O-O {-0.25/11 9} 18. Qd3 {
0.03/10 23} h6 {(Te8) -0.30/11 16} 19. a3 {(Dh3) 0.03/9 19} Kh8 {
(Te8) -0.41/11 15} 20. Qf3 {0.08/9 11} f6 {(Ld4) -0.66/12 15} 21. exf6 {
0.03/9 6} Nxf6 {-0.82/11 14} 22. Be1 {-0.18/10 19} Qc7 {-0.96/11 9} 23. Na4 {
(g3) -0.36/11 38} Ne4 {-0.95/11 7} 24. Nxc5 {-0.44/10 14} Nxc5 {-0.96/11 5} 25.
Rb6 {-0.48/11 8} Nd7 {(Se4) -1.05/13 13} 26. R6b2 {-0.48/10 10} Qxf4 {
-1.26/12 7} 27. Qxf4 {-0.54/11 15} Rxf4 {-1.36/11 8} 28. Bg3 {-0.61/12 32} Rf6
{-1.20/13 10} 29. Kg1 {(Lg4) -0.57/11 6} a5 {-1.23/13 14} 30. Re1 {
(Lg4) -0.63/10 9} Ba6 {-1.35/12 9} 31. Bf3 {(Lg4) -0.60/9 2} Kg8 {-1.27/11 9}
32. Kf2 {(Lc7) -0.67/9 14} Re8 {(Lb5) -1.95/12 11} 33. Bc7 {-1.06/10 7} e5 {
-2.08/12 7} 34. Kg3 {(Ke3) -1.07/11 5} e4 {-2.15/12 11} 35. Bh5 {-1.18/11 6} g6
{-2.30/13 11} 36. Bd1 {-1.14/12 18} a4 {(Lc4) -2.13/12 10} 37. Bf4 {-1.11/10 4}
h5 {(Kg7) -2.19/12 7} 38. Be3 {-1.16/10 8} Ne5 {(Lb5) -2.26/12 7} 39. Rb4 {
-1.22/10 4} Bb5 {-2.09/12 11} 40. Be2 {-1.05/10 3} Rf5 {(Lxe2) -2.17/11 8} 41.
Bd4 {-1.07/10 4} Bxe2 {-2.14/11 9} 42. Rxe2 {-1.19/10 4} Nd7 {(c5) -2.06/11 8}
43. Rb7 {(Txa4) -1.16/10 4} Rf7 {(c5) -2.16/11 7} 44. Rf2 {(c3) -0.83/11 4}
Rfe7 {-1.95/12 7} 45. Rc7 {-1.31/11 4} Nf8 {(Sb8) -2.29/13 7} 46. Rxc6 {
-1.47/12 5} h4+ {(e3) -2.42/13 6} 47. Kh3 {-1.56/12 4} e3 {-2.48/14 8} 48. Re2
{-1.63/13 5} Ne6 {-2.63/13 5} 49. Be5 {-1.68/13 4} Nd8 {-2.78/13 8} 50. Rxg6+ {
-1.79/13 4} Kh7 {-2.75/14 8} 51. Rg4 {-1.93/13 4} Rxe5 {-2.84/13 5} 52. Rxh4+ {
-1.74/12 2} Kg6 {-2.89/13 8} 53. Rxa4 {-1.67/12 2} Ne6 {-3.14/12 6} 54. g4 {
(g3) -1.65/13 4} Rf8 {-3.54/13 7} 55. Kg3 {-2.66/14 11} Nc5 {-3.54/13 7} 56.
Rd4 {(Tb4) -2.88/13 7} Ne4+ {-4.39/12 4} 57. Kg2 {-3.28/15 4} Nc3 {-4.76/13 6}
58. Re1 {-3.42/14 5} Rf2+ {-5.03/13 7} 59. Kh1 {-3.60/14 7} Re8 {
(Te7) -5.28/13 10} 60. Rd3 {-3.60/10 3} Rh8 {-5.28/12 6} 61. Rexe3 {
(Tdxe3) -4.38/12 6} Rhxh2+ {-5.22/9 2} 62. Kg1 {-4.20/5 0} Rfg2+ {
(Se2+) -5.27/11 5} 63. Kf1 {-3.35/5 0} Rxc2 {(Se2) -5.27/12 6} 64. Kg1 {
-5.78/13 2} Rcg2+ {-6.18/12 6} 65. Kf1 {-4.38/5 0} Ra2 {(Tf2+) -6.25/13 11} 66.
Kg1 {-4.38/12 10} Rhg2+ {-6.41/12 4} 67. Kf1 {-5.78/13 7} Raf2+ {-7.18/12 6}
68. Ke1 {-4.73/5 0} Ne4 {-7.25/13 5} 69. Rxe4 {-#21/12 2} dxe4 {-7.25/12 5} 70.
Rd6+ {(Te3) -#20/7 0} Kg5 {-#7/12 3} 71. Rd5+ {(Te6) -#19/6 0} Kh4 {
(Kxg4) -#6/15 3} 72. Rb5 {(Tc5) -#6/5 0} Ra2 {(e3) -#5/39 3} 73. Kf1 {-#5/3 0}
Kg3 {(e3) -#4/40 3} 74. Rb3+ {-#3/3 0} Kh2 {-#3/39 3} 75. Rb2 {(Th3+) -#2/3 0}
Ra1+ {-#2/35 2} 76. Rb1 {-#1/3 0} Rxb1# {-#1/34 1} 0-1



Thank You Claudio Fabiano!

[mclane]

the problem is that a match versus older versions does not really show you that there IS a progress.
it only shows that ONE engine is different than another one.

[mjlef]

Does anyone know what history depth1 means? Ryan, can you explain?

[Uri Blass]

Eduard's results was against rybka and I think that result against previous version also give important information and I think that the first test for a program should be testing against the previous version because I think that when there is a significant improvement the new version should be able to defeat the old version.

Uri

[mclane]

eduards results was blitz on a p3 600 mhz.

[Edward German]

eduards results was blitz on a p3 600 mhz.

First: Uri is right.

Next:

Then I please around tests with faster hardware, thank. It seems that this Setting is actually somewhat better than default. Test it out.

[mclane]

cool. if uri is right we don't have to discuss any further. if we have a question we only have to ask what uri says. kind of oracle from delphi

[Edward German]

cool. if uri is right we don't have to discuss any further. if we have a question we only have to ask what uri says. kind of oracle from delphi

Uri is right in the first point, test against a same program.

Next step is: Tests against another programs, and I takes here the first step against Rybka. Now should come tests against another engines, and on a faster hardware.

[Eelco de Groot]

Drawing against Rybka 1.2f is certainly not bad. I think my Gambit 006 settings are having a bit more trouble against Rybka 1.0 Beta. It is possible to try some combination of the settings, and maybe it is also possible to use Depth = 2 from Claudio also for Extended History Pruning. It certainly speeds the search up a little bit more but it may leave more holes.

So you could combine

History Depth 1 = 2 [Default =1]
History Depth 2 = 2 [*] [Default =1]
History Treshold 1 = 45 [Default = 40, 45 is a bit more aggressive History pruning]

This could still work, I have just set these options in Gambit 006 creating Gambit 007:

Result for Ivanchuk-Miton, Montreal 2007, Aufgabe 305 from Michael Gurevich:

[d]3r2k1/p1p1nppp/1p1br3/5q2/P2Pp3/BPR1P3/5PPP/1NR2QK1 b - - 6 1

Fruit 05/11/03 Gambit 007 256 Mb hash Athlon 2009 MHz

00:00:00.0 1,19 1 269 Nd5
00:00:00.0 0,00 2 701 Nd5 Rc6
00:00:00.0 1,09 2 846 c5 dxc5
00:00:00.0 0,79 3 1220 c5 dxc5 Be5 cxb6 Bxc3 Nxc3 axb6
00:00:00.0 0,79 4 2259 c5 dxc5 Be5 cxb6 Bxc3 Nxc3 axb6
00:00:00.0 0,79 5 5725 c5 dxc5 Be5 cxb6 Bxc3 Nxc3 axb6
00:00:00.0 0,79 6 15751 c5 dxc5 Be5 cxb6 Bxc3 Nxc3 axb6
00:00:00.0 0,73 7 35079 c5 dxc5 Be5 R3c2 Rh6 h3 Nd5 cxb6 Rxb6
00:00:00.1 0,73 8 73248 c5 dxc5 Be5 R3c2 Rh6 h3 Nd5 cxb6 Rxb6
00:00:00.2 0,59 9 145003 c5 dxc5 Be5 R3c2 Rh6 g3 Nd5 h4 a5 cxb6 Rxb6
00:00:00.4 0,46 10 281440 c5 dxc5 Be5 R3c2 Rh6 g3 Nd5 h4 a5 c6
00:00:00.9 0,45 11 596435 c5 dxc5 Be5 R3c2 Rh6 g3 Nd5 h4 Qg4 cxb6 Rxb6 Nd2
00:00:01.8 0,16 12 1191114 c5 dxc5 Be5 R3c2 Rh6 g3 bxc5 Bxc5 Nd5 Kh1 a5 Nd2 Rc6
00:00:08.5 0,78 13 5718690 c5 dxc5 Bxh2+ Kxh2 Rh6+ Kg1 Qh5 f4 exf3ep Qxf3 Qh2+ Kf2 Rf6 Qxf6 gxf6 cxb6 Nf5 bxa7 Qg3+ Kg1 Nxe3
00:00:17.0 0,92 14 11297159 c5 dxc5 Bxh2+ Kxh2 Rh6+ Kg1 Qh5 f4 exf3ep Qxf3 Qh2+ Kf2 Rf6 Qxf6 gxf6 cxb6 Nf5 e4 Qf4+ Kg1 Ne3 bxa7 Qxe4
00:00:25.2 0,80 15 16676478 c5 dxc5 Bxh2+ Kxh2 Rh6+ Kg1 Qh5 f4 exf3ep Qxf3 Qh2+ Kf2 Rf6 Qxf6 gxf6 cxb6 Nf5 bxa7 Qg3+ Kf1 Nh4 R3c2 Qxe3
00:01:21.9 0,68 16 57109674 c5 g3 cxd4 exd4 Qf6 Rc4 Nf5 d5 Bxa3 dxe6 Bxc1 exf7+ Qxf7 Qxc1 e3 fxe3 Nxe3 Rc8 Re8 Rxe8+ Qxe8
00:02:20.6 1,30 16 99685702 Bxh2+ Kxh2 Qh5+ Kg1 Rh6 f3 Qh2+ Kf2 Nf5 Nd2 Qh4+ Kg1 Ng3 Qd1 exf3 Qxf3 Qh2+ Kf2 Rf6 Qxf6 gxf6 Rxc7 Rd5 Rc8+ Kg7 Bf8+ Kg6 R8c7 Rf5+ Nf3
00:03:20.9 1,38 17 143582522 Bxh2+ Kxh2 Qh5+ Kg1 Rh6 f3 Nf5 Nd2 Qh1+ Kf2 Qh4+ Kg1 Ng3 Qd1 exf3 Qxf3 Qh2+ Kf2 Rf6 Qxf6 gxf6 Rxc7 Rd5 Rc8+ Kg7 Bf8+ Kg6 R8c7 Rf5+ Nf3 Ne4+ Kf1
00:05:50.3 1,58 18 252089647 Bxh2+ Kxh2 Qh5+ Kg1 Rh6 f3 Nf5 Be7 Nxe7 Rxc7 Qh2+ Kf2 exf3 Qg1 Qh4+ g3 Qf6 Rxa7 Nd5 Nc3 Nb4 Kf1 Qe6 e4 Qxb3
00:08:59.2 1,42 19 385497199 Bxh2+ Kxh2 Qh5+ Kg1 Rh6 f3 Nf5 Be7 Nxe7 Rxc7 Qh2+ Kf2 exf3 Nd2 Qh4+ g3 Qh2+ Ke1 Nd5 Nxf3 Qxg3+ Qf2 Qxf2+ Kxf2 Nxc7 Rxc7 a5 e4 Re6 Ke3
00:21:15.0 1,96 20 902209649 Bxh2+ Kxh2 Qh5+ Kg1 Rh6 f3 Nf5 Be7 Nxe7 Rxc7 Qh2+ Kf2 exf3 Nd2 Qh4+ g3 Qh2+ Ke1 Nd5 Nxf3 Qxg3+ Qf2 Rh1+ Ke2 Nf4+ exf4 Re8+ Qe3 Rxe3+ Kxe3 Rxc1 Rxc1


http://f23.parsimony.net/forum50826/messages/170132.htm

Eelco


* (I think this means something like earlier Extended History Pruning in the search, at least it means something similar in Alaric, higher numbers are more aggressive pruning, but I could be wrong?)

[Gregory Owett]

Hi,

I also tested the various Fruit versions against Strelka 1.8 with following results:

Strelka 1.8 - Fruit 2.3 5 - 5
- Fruit 06 b 5 - 5
- Fruit 05 H2 6 - 4
- Fruit 05 2.5 - 7.5

blitz (5 ' +3 ") on AMD 64 4800+ /CB /win pro 64

Apparently, the 051103 is not the weakest !

Gregory