Super Fruit (Let Secret Be Known)

[Werner]

Well, I was given permission to share settings ... lets see if my results can be duplicated .... again the tester should see that fruits evals for long time controls are very different from other engines that are out now.

...and how will you manage that?

[M ANSARI]

Rybka 3 also plays Rc4 ... but gives it a +.39 evaluation. I don't think it is "clearly winning" even when I played it out and ended up with 2 rooks vs. rook + bishop + passed pawn. The passed black pawn along with the fact that bishops are of opposite color should give good chances for a draw. A bishop and a passed pawn who happen to be of the same color are just as strong as a rook. I do agree though that white has all the winning chances. Here is Rybka 3 evaluation

New game
3r1bk1/3P2p1/6Bp/pp6/5r2/P6P/6P1/2RR3K b - - 0 1

Analysis by Rybka 3 :

1...Bxa3
³ (-0.43) Depth: 2 00:00:00
1...Bxa3
³ (-0.54) Depth: 3 00:00:00
1...Bxa3
³ (-0.43) Depth: 4 00:00:00
1...Bxa3 2.Rc8
= (-0.05) Depth: 5 00:00:00
1...Bxa3 2.Rc8 Rff8 3.Bf5
= (-0.07) Depth: 6 00:00:00 5kN
1...Bxa3
³ (-0.27) Depth: 7 00:00:00 22kN
1...Bxa3 2.Rc8 Rff8 3.Bf5 Kf7 4.Rd5
³ (-0.36) Depth: 7 00:00:00 22kN
1...Bxa3 2.Rc8 Rff8 3.Bf5 Kf7 4.Rd5
³ (-0.36) Depth: 8 00:00:00 25kN
1...Bxa3 2.Rc8 Rff8 3.Bf5 g6 4.Rxd8 Rxd8 5.Bxg6 b4 6.Bc2
³ (-0.34) Depth: 9 00:00:00 65kN
1...Rc4 2.Rb1 b4 3.axb4 axb4
² (0.28) Depth: 10 00:00:01 424kN
1...Rc4 2.Rb1 b4 3.axb4 axb4 4.Rf1 Rd4 5.Bf5 Be7 6.Be6+ Kh7 7.g3
= (0.22) Depth: 11 00:00:01 460kN
1...Rc4 2.Rb1 b4 3.axb4 axb4 4.Rf1 Rd4 5.Bf5 Be7 6.Be6+ Kh7 7.g3
= (0.22) Depth: 12 00:00:01 477kN
1...Rc4 2.Rb1 b4 3.axb4 axb4 4.Rf1 Rd4 5.Bf5 Be7 6.Be6+ Kh7 7.g3 Kg6 8.Rf2
= (0.20) Depth: 13 00:00:02 610kN
1...Rc4 2.Rb1 b4 3.axb4 axb4 4.Rf1 Rd4 5.Bf5 Be7 6.Be6+ Kh7 7.g3 Kg6 8.Rf2
= (0.20) Depth: 14 00:00:02 668kN
1...Rc4 2.Rb1 Rc7 3.Be8 b4 4.axb4 axb4 5.Rf1 Bd6 6.h4 Ra7 7.h5 Ra5 8.Bf7+
= (0.25) Depth: 15 00:00:03 1348kN
1...Rc4 2.Rb1 Rc7 3.Be8 b4 4.axb4 axb4 5.Rf1 Be7 6.Rfe1 Bf8 7.Re5 Rc6 8.Re2 g5 9.Bh5 Rxd7 10.Be8 Rdc7 11.Bxc6
² (0.28) Depth: 16 00:00:05 2153kN
1...Rc4 2.Rb1 Rc7 3.Be8 b4 4.axb4 axb4 5.Rf1 Be7 6.Rfe1 Bf8 7.Re5 Rc6 8.Re2 g5
² (0.28) Depth: 17 00:00:09 3603kN
1...Rc4 2.Rb1 Rc7 3.Be8 b4 4.axb4 axb4 5.Rf1 Be7 6.Rfe1 Bf8 7.Re5 Rc6 8.Re2 g5
² (0.28) Depth: 18 00:00:10 4038kN
1...Rc4 2.Rb1 Rc7 3.Be8 b4 4.axb4 axb4 5.Rf1 Be7 6.Rfe1 Bf8 7.Re2 g5 8.g4 Rc3 9.Kg2
² (0.33) Depth: 19 00:00:27 11776kN
1...Rc4 2.Rb1 Rc7 3.Be8 b4 4.axb4 axb4 5.Rf1 Be7 6.Rfe1 Rdxd7 7.Bxd7 Rxd7 8.g3 Kf7 9.Rb3 Ra7 10.Rf3+ Kg6 11.Re6+ Bf6 12.Rb6 Ra1+ 13.Kg2 Rb1
² (0.35) Depth: 20 00:01:03 26915kN
1...Rc4 2.Rb1 Rc7 3.Be8 b4 4.axb4 axb4 5.Rf1 Be7 6.Rfe1 Rdxd7 7.Bxd7 Rxd7 8.g3 Kf7 9.Rb3 Ra7 10.Rf3+ Kg6 11.Re6+ Bf6 12.Rb6 Ra1+ 13.Kg2 Rb1
² (0.35) Depth: 21 00:01:20 34429kN
1...Rc4 2.Rb1 Rc7 3.Be8 b4 4.axb4 axb4 5.Rf1 Be7 6.Rfe1 Rdxd7 7.Bxd7 Rxd7 8.g3 Kf7 9.Rb3 Ra7 10.Rf3+ Kg6 11.Re6+ Bf6 12.Rb6 Ra1+ 13.Kg2 Rb1
² (0.35) Depth: 22 00:01:30 38777kN
1...Rc4 2.Rb1 Rc7 3.Be8 b4 4.axb4 axb4 5.Rf1 Be7 6.Rfe1 Rdxd7 7.Bxd7 Rxd7 8.g3 Kf7 9.Rb3 Ra7 10.Rf3+ Kg8 11.Kg2 Bc5 12.Re8+ Kh7 13.Re2 Kg6 14.Re6+
² (0.39) Depth: 23 00:03:29 94760kN

(, 05.12.2008)

[Werner]

Yes what you are saying is true. I guess I should have said white is no losing. Some have posted that black is up by .24


About this version and settings, can you post the settings you have. Reason I asked .... I dont know of anyone that has this version. It was a special version with certain option that I wanted to do adjusting out side of the options that others had.

But, if its ok with Ryan and Fabien I will release the settings.

Hi,
I am quite sure (and so is Ryan) that´s not possible to raise an engines elo for over 200 only with changing the settings.
At the moment in CEGT Rybka is around 240 elos above of each Fruit version we tested (and 170 above the next one)