Fruit 05/11/03 tourney

[Gandalf]

This is a 64 bit compile (optimized for Intel Chips) created with the Orcas Beta compiler from Micosoft. You fill find it a shade slower then the 32 bit compile.

Wait... Slower? What's the point then?

[VP]

Fruit23-EM64T (what is this?)

This is a 64 bit compile (optimized for Intel Chips) created with the Orcas Beta compiler from Micosoft. You fill find it a shade slower then the 32 bit compile.


Bryan

So what does it mean?
Since it is a 64 bit compile, would it be stronger because it is taking advantage of more virtual memory available?
My 64 bit Core2Duo T7300 runs 32 bit engines also.
Would it be better if I simply ran the 32 bit Fruit 2.3?
Thanks for your time.

[Dirt]

This is a 64 bit compile (optimized for Intel Chips) created with the Orcas Beta compiler from Micosoft. You fill find it a shade slower then the 32 bit compile.

Wait... Slower? What's the point then? :?

It probably has to do with the fact that if he doesn't provide the slower 64 bit compiles he is constantly nagged by requests for them. Besides, maybe when the new processors expected soon from both AMD and Intel come out the 64 bit versions will show a speed up.

[Bryan Hofmann]

This is a 64 bit compile (optimized for Intel Chips) created with the Orcas Beta compiler from Micosoft. You fill find it a shade slower then the 32 bit compile.

Wait... Slower? What's the point then?

It probably has to do with the fact that if he doesn't provide the slower 64 bit compiles he is constantly nagged by requests for them. Besides, maybe when the new processors expected soon from both AMD and Intel come out the 64 bit versions will show a speed up.

BINGO give this man the grand prize. I am always asked for a 74 bit compile by end users and have to go thru the same explaination every time as to why I did not create one. So I made one this time so the users can see for themselves. The difference in speed will make zero difference.


Bryan

[ozziejoe]

fruit also lost to toga in my tourni, 11-17-34....However, it beat shredder 10 pretty handily...30-19-13.. (am going on memory as am currently testing fruit against rybka mp)


It looks like a sizeable improvement for sure.


I don't recollect seeing a fruit 2.3 with the file i downloaded? is there a newer file.

[Eelco de Groot]

fruit also lost to toga in my tourni, 11-17-34....However, it beat shredder 10 pretty handily...30-19-13.. (am going on memory as am currently testing fruit against rybka mp)


It looks like a sizeable improvement for sure.


I don't recollect seeing a fruit 2.3 with the file i downloaded? is there a newer file.

Hello Joseph, yes maybe you missed the last 5 new versions ?
http://www.talkchess.com/forum/viewtopic.php?t=15358

I'm still testing the 05/11/03 version of Fruit and having some problems when changing the UCI-options at least in the Shredder interface.

Example position:


[d]4k2r/2qb1pb1/3ppnpp/8/1R2P1PP/2N1BP2/1PPQ4/NK5R b k -

Engine: HIARCS 11.1 UCI (128 MB)
gemaakt door Mark John Uniacke

8/23 0:00 -7.32-- 22...Kf8 (70.008) 131

8/23 0:00 -7.32 22...Kf8 23.Td1 Pe8 24.Pb5 Lxb5
25.Txb5 h5 (87.743) 143

9/23 0:01 -7.32 22...Kf8 23.Td1 Pe8 24.Pb5 Lxb5
25.Txb5 h5 26.gxh5 Txh5 27.Txh5 gxh5
28.Lf4 Ke7 (185.088) 137

10/23 0:03 -7.45 22...Kf8 23.Pb3 Pe8 24.Pb5 Lxb5
25.Txb5 Th7 26.Pd4 Dc4 27.Tb7 Pc7
28.Tb8+ (460.451) 142

10/23 0:04 -7.44 22...Ke7 (599.690) 142

10/26 0:05 -7.44 22...Ke7 23.Pb3 (783.718) 144

10/27 0:07 -7.43 22...Lf8 (1.066.757) 139

10/27 0:07 -7.43 22...Lf8 23.Pb5 Dc6 24.Tb3 Th7
25.Td1 (1.077.616) 139

11/28 0:09 -7.66 22...Lf8 23.Pb5 Lxb5 24.Txb5 (1.333.582) 139

11/28 0:09 -7.65 22...Ke7 (1.406.474) 141

11/29 0:12 -7.60 22...Ke7 23.Pb3 Pe8 24.Ld4 Lf6
25.Lxf6+ (1.805.716) 145

11/29 0:13 -7.60 22...Kf8 (1.900.696) 146

11/29 0:13 -7.48 22...Kf8 23.Pb3 Pe8 24.Pb5 Lxb5
25.Txb5 Th7 26.Pd4 h5 27.gxh5 Txh5
28.Txh5 gxh5 29.Tg1 Dc4 (1.968.779) 146

12/29 0:20 -7.64 22...Kf8 23.Pb3 Kg8 24.Td1 Pe8
25.Pb5 Lxb5 26.Txb5 h5 27.gxh5 Txh5
28.Txh5 (2.986.017) 146

12/29 0:23 -7.64 22...Ke7 (3.550.573) 151

12/29 0:24 -7.61 22...Ke7 23.Pb3 Pe8 24.Ld4 Kf8 25.h5 Lxd4
26.Txd4 Kg7 27.hxg6 fxg6 28.f4 (3.749.679) 151

13/34 0:47 -7.85 22...Ke7 23.Pb3 Pe8 24.Pb5 Dc8
25.Td1 Lxb5 26.Txb5 (7.488.231) 157

13/34 0:51 -7.84 22...Kf8 (8.014.711) 156

13/34 0:54 -7.65 22...Kf8 23.Pb3 Kg8 24.Td1 Pe8
25.Lf4 e5 (8.462.419) 156

14/34 1:21 -7.64 22...Kf8 23.Pb3 Kg8 24.Td1 Pe8
25.Pb5 Lxb5 26.Txb5 Dd8 27.Lf2 h5
28.gxh5 Df6 29.Pd4 Txh5 30.Txh5 (12.324.826) 151

beste zet: Ke8-f8 tijd: 1:55.297 min n/s: 146.802 nodes: 16.914.616


I thought I had it fixed, by changing Material from 102 back to 100 and then to 102 again in the settings, the eval of the following position was correct again then, but after reloading the engine with my slightly changed settings something still seems to be broken in a new match, Fruit gives away all its material as quickly as possible, and the evaluation seems not to take material into account:

4k2r/2qb1pb1/3ppnpp/8/1R2P1PP/2N1BP2/1PPQ4/NK5R b k -

Engine: Fruit 05/11/03 Beta Gambit 005B (128 MB)
gemaakt door Fabien Letouzey

1/10 0:00 0.00 22...Pxe4 23.fxe4 (161)

2/10 0:00 -0.16 22...Pxe4 23.Tb8+ Dxb8 24.fxe4 (306)

3/10 0:00 -0.10 22...Pxe4 23.fxe4 Ld4 24.Pb3 (1.511)

4/13 0:00 -0.09 22...Pxe4 23.Pb3 Pxd2+ 24.Pxd2 Ld4
25.Lxd4 (5.114)

5/16 0:00 -0.14 22...Pxe4 23.Tb8+ Ke7 24.fxe4 Txb8
25.Tf1 Lxc3 (17.403)

5/16 0:00 -0.10 22...O-O 23.Pb3 Pxe4 24.fxe4 Lxc3
25.Pc5 (21.305)

6/16 0:00 -0.03 22...O-O 23.Pb3 Pxe4 24.fxe4 Lxc3
25.Dxc3 Dxc3 26.Tb7 (36.716)

7/18 0:00 -0.03 22...O-O 23.Pb3 Pxe4 24.fxe4 Lxc3
25.Tb7 Ta8 26.Txc7 (79.010)

8/20 0:00 -0.02 22...O-O 23.Pb3 Pxe4 24.fxe4 Lxc3
25.Tb7 Ta8 26.Txc7 Ta1+ 27.Kxa1 (190.162)

9/22 0:01 -0.15 22...O-O 23.Pb3 Pxe4 24.fxe4 Lxc3
25.Tb7 Ta8 26.Txc7 Ta2 27.Kxa2 (411.041) 374

10/29 0:05 -0.04 22...O-O 23.Pb3 Ta8 24.Pc5 Ta2
25.Kxa2 Pxe4 26.P5xe4 Ld4 27.Lxd4 Da5+
28.Kb1 (2.229.219) 376

11/29 0:10 -0.06 22...O-O 23.Pb3 Ta8 24.Pc5 Pxe4
25.fxe4 Ta2 26.Kxa2 Lb5 27.Ta4 Db7
28.Kb1 (4.014.933) 378

12/29 0:26 -0.13 22...O-O 23.Pb3 Ta8 24.h5 Pxe4
25.Pxe4 d5 26.hxg6 fxg6 27.Tb8+ Dxb8
28.Pd4 e5 29.Pe6 Dxb2+ 30.Kxb2 Tb8+
31.Ka2 (10.268.737) 381

13/34 1:01 -0.19 22...O-O 23.Pb3 Ta8 24.h5 Pxe4
25.Pxe4 g5 26.Dxd6 Dxd6 27.Pxd6 Lb5
28.Txb5 Ta1+ 29.Kxa1 Le5 30.Txe5 (23.279.503) 381

beste zet: O-O tijd: 1:55.157 min n/s: 386.659 CPU 100.0% n/s(1CPU): 386.659 nodes: 44.460.000

Closing down Shredder, change to a higher resolution so I can see all the settings, reload the engine and maby because I have not yet tried to play a match, now the eval is as it should be again, using the same Gambit 005B settings:

4k2r/2qb1pb1/3ppnpp/8/1R2P1PP/2N1BP2/1PPQ4/NK5R b k -

Engine: Fruit Beta Gambit 005B (128 MB)
gemaakt door Fabien Letouzey

1/09 0:00 -6.00 22...Lc6 (145)

2/09 0:00 -6.15 22...Lc6 23.Pb3 (262)

2/09 0:00 -6.14 22...O-O 23.Lxh6 (368)

2/09 0:00 -5.71 22...h5 23.Lf4 (583)

3/11 0:00 -6.19 22...h5 23.g5 Ph7 (1.236)

4/11 0:00 -6.21 22...h5 23.g5 Ph7 24.Pb3 Lxc3
25.Dxc3 Dxc3 26.bxc3 (2.745)

5/13 0:00 -6.34 22...h5 23.g5 Ph7 24.Ld4 e5 (5.497)

6/13 0:00 -6.24 22...h5 23.g5 Ph7 24.Ld4 e5 25.Pd5 (12.648)

7/14 0:00 -6.68 22...h5 23.g5 Ph7 24.Ld4 O-O 25.Lxg7 Kxg7
26.Pb3 (24.534)

7/19 0:00 -6.61 22...Ph7 23.Td1 g5 24.Pb3 Lxc3
25.Dxc3 Dxc3 26.bxc3 Ke7 (88.380)

7/19 0:00 -6.51 22...Pg8 23.Ld4 Pf6 24.f4 O-O 25.Lb6 Db7 (98.620)

8/19 0:00 -6.43 22...Pg8 23.Ld4 Pf6 24.f4 O-O 25.g5 Pg4
26.Lxg7 Kxg7 (129.174)

9/20 0:00 -6.92 22...Pg8 23.Ld4 Pf6 24.Lb6 Dc6
25.Pb3 Lc8 26.Td1 h5 (195.607)

9/22 0:00 -6.71 22...h5 23.g5 Pg8 24.Ld4 e5 25.Pd5 Dc6
26.Tb6 Dc8 (306.631)

9/23 0:00 -6.55 22...Lf8 23.Lb6 Dc8 24.Te1 d5 25.Tb3 dxe4
26.Dd4 Le7 27.Pxe4 (430.446)

10/26 0:01 -6.44 22...Lf8 23.Lb6 Dc8 24.Pb3 d5 25.Td4 Da6
26.Lc7 Da7 27.Lf4 (872.758) 610

11/27 0:02 -6.74 22...Lf8 23.Td1 Le7 24.Pb5 Lxb5
25.Txb5 Pd7 26.Ta5 Dd8 27.Dd4 Lf6
28.Db4 (1.507.234) 605

12/29 0:04 -6.82 22...Lf8 23.Td1 Le7 24.Pb5 Lxb5
25.Txb5 Pd7 26.Ta5 Lxh4 27.Ta8+ Ld8
28.Ta7 Dc6 29.Dxd6 Dxd6 30.Txd6 (2.972.662) 615

13/31 0:09 -6.95 22...Lf8 23.Td1 Le7 24.Pb5 Lxb5
25.Txb5 h5 26.g5 Pd7 27.Ta5 O-O
28.Ta7 Dc6 29.Pb3 (6.094.051) 633

14/35 0:24 -7.24 22...Lf8 23.Td1 Le7 24.Lxh6 Lc6
25.Lg5 O-O 26.Pb3 Tc8 27.Pd4 Le8
28.h5 d5 29.Lf4 (15.798.060) 640

14/43 1:03 -7.19 22...Kf8 23.Pb3 Pe8 24.Pd4 Kg8
25.Dd3 h5 26.g5 e5 27.Pb3 Dd8 28.Tb7 Kh7
29.Pd5 (41.694.502) 654

15/43 1:29 -7.18 22...Kf8 23.Pb3 Kg8 24.Tb6 Pe8
25.Pb5 Lxb5 26.Txb5 h5 27.g5 Kh7
28.Pd4 Lxd4 29.Lxd4 e5 30.Lb6 Dc4 (58.655.571) 654

beste zet: Ke8-f8 tijd: 1:33.125 min n/s: 655.796 CPU 100.0% n/s(1CPU): 655.796 nodes: 61.010.000

Grrrr.. I'll try again whether a match now works.

Regards, Eelco

[Eelco de Groot]

No, in a match Fruit still plays Santa, what is this version called on ICC agin where you have to give away as many pieces as possible? But Hiarcs does not want to play along...


[Event "6 minuten per partij + 6 seconden per zet"]
[Site "Engine Match"]
[Date "2007.07.28"]
[Round "1"]
[White "Fruit Beta Gambit 005B"]
[Black "HIARCS 11.1 UCI"]
[Result "0-1"]

1. d4 {Boek 0s} Nf6 {Boek 0s} 2. Nf3 {Boek 0s} g6 {Boek 0s}
3. g3 {Boek 0s} Bg7 {Boek 0s} 4. Bg2 {Boek 0s} O-O {Boek
0s} 5. O-O {Boek 0s} d6 {Boek 0s} 6. c4 {Boek 0s} c6 {Boek
0s} 7. Nc3 {Boek 0s} Qa5 {Boek 0s} 8. e4 {Boek 0s} Bg4
{Boek 0s} 9. h3 {Boek 0s} Bxf3 {-0.63/13 22s} 10. Bxf3
{Boek 0s} Nbd7 {Boek 0s} 11. Qe2 {Boek 0s} e5 {Boek 0s}
12. d5 {Boek 0s} cxd5 {-0.62/13 17s} 13. Rd1 {0.00/12 17s}
d4 {+0.66/14 17s} 14. Rxd4 {-0.03/12 20s} exd4 {+5.92/13
10s} 15. Be3 {-0.07/12 21s} dxc3 {+9.40/12 10s} 16. c5
{-0.04/12 30s} cxb2 {+11.08/13 16s} 17. Qxb2 {-0.07/11 16s}
Ng4 {+12.12/13 16s} 18. cxd6 {0.00/11 16s} Bxb2 {+20.89/14
11s} 19. Bxg4 {-0.19/12 13s} Bxa1 {+21.35/14 10s} 20. Kg2
{-0.43/13 21s} Ne5 {+25.59/13 12s} 21. a4 {-0.34/12 14s}
Nxg4 {+28.92/15 15s} 22. hxg4 {-1.43/14 3:57m} Qxa4
{+29.60/14 15s} 23. f3 {-1.19/13 10s} Qc2+ {+M8/15 4s}
24. Kh3 {-M7/12 10s} Qe2 {+M7/13 1s} 25. g5 {-M5/23 5s}
Qf1+ {+M6/11 0s} 26. Kg4 {-M5/33 5s} f5+ {+M5/9 0s}
27. gxf6 {-M4/45 5s} Qh1 {+M4/7 0s} 28. Bg1 {-M3/49 6s}
Rxf6 {+M3/5 0s} 29. Bh2 {-M2/47 5s} Qxh2 {+M2/3 0s} 30. f4
{-M1/45 5s} Qh5# {+M1/1 0s} 0-1

[d]r5k1/pp5p/3P1rp1/7q/4PPK1/6P1/8/b7 w - -

[Event "6 minuten per partij + 6 seconden per zet"]
[Site "Engine Match"]
[Date "2007.07.28"]
[Round "2"]
[White "HIARCS 11.1 UCI"]
[Black "Fruit Beta Gambit 005B"]
[Result "1-0"]

1. d4 {Boek 0s} Nf6 {Boek 0s} 2. Nf3 {Boek 0s} g6 {Boek 0s}
3. g3 {Boek 0s} Bg7 {Boek 0s} 4. Bg2 {Boek 0s} O-O {Boek
0s} 5. O-O {Boek 0s} d6 {Boek 0s} 6. c4 {Boek 0s} c6 {Boek
0s} 7. Nc3 {Boek 0s} Qa5 {Boek 0s} 8. e4 {Boek 0s} Bg4
{Boek 0s} 9. h3 {Boek 0s} c5 {+0.22/12 33s} 10. hxg4
{+3.46/13 30s} Nc6 {+0.30/12 27s} 11. d5 {+3.99/13 18s} Nd4
{+0.35/11 18s} 12. Nxd4 {+4.03/13 17s} cxd4 {+0.45/12 21s}
13. Nb5 {+4.00/13 14s} Rac8 {+0.57/13 19s} 14. b3 {+4.20/13
20s} Rxc4 {+0.82/12 17s} 15. bxc4 {+8.73/12 10s} Nxd5
{+0.69/12 16s} 16. exd5 {+11.75/13 10s} Be5 {+0.51/13 34s}
17. Qd3 {+12.73/13 15s} Rc8 {+0.49/12 14s} 18. Bg5
{+12.94/13 13s} a6 {+0.26/12 35s} 19. Nxd4 {+12.97/13 9s}
Qc5 {+0.20/12 12s} 20. Rfd1 {+13.05/13 15s} Re8 {+0.12/12
28s} 21. Rab1 {+13.52/14 14s} b5 {-0.08/12 7s} 22. cxb5
{+15.11/12 9s} axb5 {-0.24/13 17s} 23. Nc6 {+15.42/13 14s}
Qc3 {-0.16/11 14s} 24. Qxb5 {+18.17/12 9s} h6 {-0.60/12
21s} 25. Nxe7+ {+27.00/12 9s} Rxe7 {-1.02/13 8s} 26. Bxe7
{+30.83/14 10s} Qc8 {-8.31/16 2:38m} 27. Rdc1 {+34.57/14
13s} Qa8 {-1.24/13 7s} 28. Qd7 {+M7/13 6s} Bc3 {-M6/10 5s}
29. Rxc3 {+M6/11 0s} Kh7 {-M5/15 3s} 30. Bf6 {+M5/9 0s} Qf8
{-M4/45 4s} 31. Rc8 {+M4/7 0s} g5 {-M3/44 4s} 32. Rxf8
{+M3/5 0s} h5 {-M2/44 4s} 33. gxh5 {+M2/3 0s} g4 {-M1/44
5s} 34. Rh8# {+M1/1 0s} 1-0

[d]7R/3Q1p1k/3p1B2/3P3P/6p1/6P1/P4PB1/1R4K1 b - -

Eelco

[Eelco de Groot]

I tried a whole new installation of Shredder on my D: partition, suspecting maybe something was not right with the Shredder GUI or maybe I had too many engines installed, but the problem with a complete reinstallation is still there.

I can completely reproduce the analysis every time and the differences remain the same. Analysis after loading the engines directly into the GUI is correct, but going to tournament mode, loading the engines, then cancelling the tournament or finish a tournament, in Analysis mode the output is different and in the case of some parameter changes completely nonsensical. The engine still correctly sees matescores but for instance does not take into account material anymore if any changes are made to the Material parameter, for instance just increasing it to 102. Eval should be around -20 at the end of a game but is around zero. Because the same wrong settings are used for a match, any match results are without relation to the chosen settings. So any testing of the settings in matches this way becomes impossible.

For the default values there seems to be no difference how the engine is loaded into the GUI.

An example in the same position from previous post, here with slightly different settings not using the disastrous Material parameter anymore, the output is a little better , -6.7 instead of 0.00 and better moves are chosen, but in case the engine is loaded in tournament mode still not the way the settings should work and the tuning is off.

After loading the engines in newly installed GUI but not using Tournament mode:


[d]4k2r/2qb1pb1/3ppnpp/8/1R2P1PP/2N1BP2/1PPQ4/NK5R b k -

Engine: Fruit 05/11/03 (128 MB)
by Fabien Letouzey

1/08 0:00 -5.66 22...Bc6 (138)

2/08 0:00 -5.82 22...Bc6 23.Nb3 (254)

2/08 0:00 -5.81 22...O-O 23.Bxh6 (362)

2/08 0:00 -5.38 22...h5 23.Bf4 (570)

3/11 0:00 -6.00 22...h5 23.g5 Nh7 (1.216)

4/11 0:00 -6.06 22...h5 23.g5 Nh7 24.Nb3 Bxc3
25.Qxc3 Qxc3 26.bxc3 (2.726)

4/11 0:00 -6.01 22...e5 23.Bb6 Qb8 24.Nb3 (4.180)

5/13 0:00 -6.41 22...e5 23.Nb3 Bc6 24.Nb5 Bxb5
25.Rxb5 (9.873)

5/13 0:00 -6.10 22...h5 23.g5 Nh7 24.Bd4 e5 (11.333)

6/13 0:00 -6.05 22...h5 23.g5 Ng8 24.Bd4 e5 25.Nd5 (18.292)

7/15 0:00 -6.47 22...h5 23.g5 Nh7 24.Bd4 O-O 25.Bxg7 Kxg7
26.Nb3 (34.274)

7/20 0:00 -6.31 22...Bc6 23.Nb3 Nd7 24.Nd4 Ke7
25.Ncb5 Bxb5 26.Rxb5 (56.077)

8/20 0:00 -6.24 22...Bc6 23.Nb3 Nd7 24.Nd4 h5
25.Nxc6 Qxc6 26.Bd4 Bxd4 27.Rxd4 (122.581)

8/23 0:00 -6.18 22...Bf8 23.Nb3 d5 24.Rd4 dxe4
25.Nxe4 Nxe4 26.Rxe4 Bc6 27.Qc3 (177.548)

9/23 0:00 -6.36 22...Bf8 23.Nb3 d5 24.Rd4 Qg3 25.Qe2 dxe4
26.Nxe4 Nxe4 27.Rxe4 Be7 (240.310)

10/26 0:01 -6.62 22...Bf8 23.Nb5 Bxb5 24.Rxb5 d5
25.Ra5 Qd8 26.Rd1 Be7 27.Qc3 Bd6
28.exd5 exd5 (613.691) 325

10/26 0:02 -6.52 22...h5 23.g5 Ng8 24.Bd4 e5 25.Nd5 Qc6
26.Rb8+ Bc8 27.Nb6 Ne7 28.Nxc8 Nxc8 (930.001) 315

10/26 0:03 -6.46 22...O-O 23.Bf4 Ra8 24.Bxd6 Qc6
25.Nb3 h5 26.Rg1 hxg4 27.fxg4 Nd5
28.exd5 Qxd6 29.dxe6 Qxe6 (1.216.435) 316

11/28 0:07 -7.38 22...O-O 23.Bxh6 Ra8 24.Bxg7 Kxg7
25.Nb3 Kf8 26.Qh6+ Ke7 27.Rd1 Bc6
28.Qf4 Bb7 (2.355.815) 321

11/29 0:07 -6.74 22...h5 23.g5 Ng8 24.Bd4 Bxd4
25.Rxd4 Ne7 26.Rd1 Nc8 27.Nb3 Bc6
28.Qf4 (2.553.961) 321

11/29 0:09 -6.51 22...Bf8 23.Nb5 Bxb5 24.Rxb5 h5
25.Qb4 Nd7 26.Nb3 Qc8 27.gxh5 d5
28.Qa4 (2.907.817) 321

12/29 0:12 -6.70 22...Bf8 23.Nb5 Bxb5 24.Rxb5 h5
25.Qd4 Be7 26.g5 e5 27.Qa4 Nd7
28.Nb3 O-O (4.018.980) 322

13/37 0:32 -6.83 22...Bf8 23.Rb6 e5 24.Nb3 Be7 25.Ra6 Qb8
26.Bxh6 Qb7 27.Ra5 Bd8 28.Ra3 Bc7 (10.496.349) 326

14/41 1:17 -6.72 22...Bf8 23.Rb6 Be7 24.Nb3 Bc6
25.Ra6 Qc8 26.Ra7 h5 27.g5 Nd7
28.Na5 Ba8 29.Nc4 (25.487.461) 329

15/41 1:53 -6.91 22...Bf8 23.Rb6 Be7 24.Nb3 Bc6
25.Ra6 Qc8 26.Ra7 h5 27.g5 Nd7
28.Rd1 Rg8 29.Bd4 Qb8 (37.760.762) 331

15/46 3:18 -6.81 22...Kf8 23.Nb3 Ne8 24.Qd3 h5 25.g5 Bxc3
26.Rc4 Bb5 27.Rxc7 Bxd3 28.Rxc3 Ba6
29.Rc6 Bb5 (66.537.362) 334

16/46 4:48 -6.94 22...Kf8 23.Nb3 Ne8 24.Nd4 h5 25.g5 Kg8
26.Rd1 Kh7 27.Rb3 Qa5 28.Rb7 Qd8
29.Ra7 Nc7 30.Qh2 (96.581.040) 334

17/46 10:01 -6.98 22...Kf8 23.Nb3 Kg8 24.Rd1 Bf8
25.Nb5 Bxb5 26.Rxb5 d5 27.exd5 Nxd5
28.h5 gxh5 29.gxh5 Nxe3 30.Qxe3 Be7
31.f4 Kh7 32.Qe4+ Kg7 33.Rg1+ Kf8 (201.566.940) 335


4k2r/2qb1pb1/3ppnpp/8/1R2P1PP/2N1BP2/1PPQ4/NK5R b k -

Engine: Fruit 05/11/03 Gambit 005 [new] (128 MB)
by Fabien Letouzey

1/09 0:00 -5.94 22...Bc6 (145)

2/09 0:00 -6.09 22...Bc6 23.Nb3 (262)

2/09 0:00 -6.08 22...O-O 23.Bxh6 (368)

2/09 0:00 -5.65 22...h5 23.Bf4 (583)

3/11 0:00 -6.13 22...h5 23.g5 Nh7 (1.238)

4/11 0:00 -6.11 22...h5 23.g5 Nh7 24.Nb3 Bxc3
25.Qxc3 Qxc3 26.bxc3 (2.765)

5/13 0:00 -6.28 22...h5 23.g5 Nh7 24.Bd4 e5 (5.433)

6/13 0:00 -6.18 22...h5 23.g5 Nh7 24.Bd4 e5 25.Nd5 (11.842)

7/16 0:00 -6.61 22...h5 23.g5 Nh7 24.Bd4 O-O 25.Bxg7 Kxg7
26.Nb3 (24.710)

7/24 0:00 -6.47 22...Ng8 23.Bd4 Nf6 24.Nb3 O-O 25.g5 hxg5
26.hxg5 (85.296)

8/24 0:00 -6.50 22...Ng8 23.Bd4 Nf6 24.Nb3 e5 25.Bb6 Qc6
26.Rd1 (107.631)

8/24 0:00 -6.32 22...Bf8 23.Bb6 Qc8 24.Nb3 d5 25.Rd4 Qa6
26.Bc5 (171.194)

9/24 0:00 -6.33 22...Bf8 23.Bb6 Qc6 24.Rd1 d5
25.exd5 exd5 26.Rb5 Be7 (279.841)

10/27 0:02 -6.69 22...Bf8 23.Nb5 Bxb5 24.Rxb5 d5
25.Ra5 Qd8 26.exd5 Nxd5 27.Bd4 Rg8
28.Nb3 (660.627) 302

11/31 0:05 -6.62 22...Bf8 23.Nb5 Bxb5 24.Rxb5 h5
25.Ra5 Qc6 26.Qb4 hxg4 27.Qb8+ Ke7
28.Bg5 Qd7 29.Bxf6+ Kxf6 30.fxg4 (1.798.035) 313

12/31 0:08 -6.77 22...Bf8 23.Nb5 Bxb5 24.Rxb5 h5
25.Ra5 Qc6 26.g5 Nd7 27.Rd1 Be7
28.Nb3 Ne5 (2.747.421) 316

13/31 0:17 -6.87 22...Bf8 23.Nb5 Bxb5 24.Rxb5 h5
25.Ra5 Qc6 26.g5 Nd7 27.Rd1 Be7
28.Nb3 Ne5 29.f4 (5.658.594) 322

14/35 0:40 -7.06 22...Bf8 23.Nb5 Bxb5 24.Rxb5 Be7
25.Ra5 Qc6 26.Nb3 d5 27.Qh2 Bd6
28.e5 Bc7 29.Ra7 O-O 30.Bxh6 (13.438.182) 328

14/38 1:21 -7.05 22...Kf8 23.Nb3 Ne8 24.Nd4 h5 25.g5 Qc8
26.Rd1 e5 27.Nb3 Kg8 28.Nd5 Qa6
29.Qg2 (27.100.058) 330

15/47 2:51 -6.97 22...Kf8 23.Nb3 Kg8 24.Qd3 Bf8
25.Rb6 Kg7 26.Ra6 Be7 27.Ra7 Qc6
28.h5 g5 29.Nd4 Qc5 (56.299.445) 328

16/47 4:39 -7.17 22...Kf8 23.Nb3 Kg8 24.Qd3 h5 25.Rc4 Qd8
26.g5 Ne8 27.Rd1 Qb8 28.f4 e5 29.Nd5 exf4
30.Bxf4 Qa7 31.Ne7+ Kh7 32.Bxd6 (93.000.696) 332

17/51 10:04 -7.18 22...Kf8 23.Nb3 Ne8 24.Nd4 Kg8
25.Rb3 h5 26.g5 Kh7 27.Ndb5 Qc6
28.Rd1 f5 29.gxf6 Bxf6 30.Nxd6 Nxd6
31.Qxd6 Qxd6 32.Rxd6 (202.616.877) 334

(This is the desired output)

And after loading the engines in Tournament mode, then cancelling the
tournament, the analysis should be the same but is not, for any changed
parameter settings:


4k2r/2qb1pb1/3ppnpp/8/1R2P1PP/2N1BP2/1PPQ4/NK5R b k -

Engine: Fruit 05/11/03 Gambit 005 [new] (128 MB)
by Fabien Letouzey

1/09 0:00 -5.80 22...Bc6 (145)

2/09 0:00 -5.95 22...Bc6 23.Nb3 (260)

2/09 0:00 -5.94 22...O-O 23.Bxh6 (366)

2/09 0:00 -5.51 22...h5 23.Bf4 (574)

3/11 0:00 -6.13 22...h5 23.g5 Nh7 (1.224)

4/11 0:00 -6.11 22...h5 23.g5 Nh7 24.Nb3 Bxc3
25.Qxc3 Qxc3 26.bxc3 (2.780)

5/13 0:00 -6.28 22...h5 23.g5 Nh7 24.Bd4 e5 (5.366)

6/13 0:00 -6.18 22...h5 23.g5 Nh7 24.Bd4 e5 25.Nd5 (13.073)

7/15 0:00 -6.61 22...h5 23.g5 Nh7 24.Bd4 O-O 25.Bxg7 Kxg7
26.Nb3 (26.556)

7/24 0:00 -6.50 22...Bc6 23.Rd1 Bf8 24.Nb3 d5 25.Nc5 Bxc5
26.Bxc5 (53.755)

7/24 0:00 -6.32 22...Ng8 23.Nb3 Bxc3 24.Qxc3 Qxc3
25.bxc3 Ne7 26.Rd1 Nc6 (111.857)

8/24 0:00 -6.33 22...Ng8 23.Nb3 Bxc3 24.Qxc3 Qxc3
25.bxc3 Ke7 26.Rb8 h5 27.gxh5 (138.226)

8/24 0:00 -6.18 22...Bf8 23.Bb6 Qc8 24.Nb3 d5 25.Rd4 Qa6
26.Bc5 (208.543)

9/24 0:01 -6.32 22...Bf8 23.Nb5 Bxb5 24.Rxb5 h5
25.g5 Nd7 26.Nb3 Ne5 27.Qe2 (325.017) 300

10/26 0:01 -6.75 22...Bf8 23.Nb5 Bxb5 24.Rxb5 Nd7
25.Nb3 Bg7 26.Ra5 h5 27.Ra8+ Nb8
28.Bd4 Bxd4 29.Qxd4 (603.482) 302

10/26 0:02 -6.52 22...Ng8 23.Nb3 Bxc3 24.Qxc3 Qxc3
25.bxc3 Ke7 26.Rb8 f6 27.Rd1 g5
28.h5 (783.855) 305

11/33 0:05 -6.83 22...Ng8 23.Bb6 Bxc3 24.bxc3 Qc6
25.Nb3 Ne7 26.Rd1 Kf8 27.Qxd6 Qxc3
28.Bd4 (1.799.932) 320

11/36 0:13 -6.82 22...d5 23.exd5 O-O 24.dxe6 Bxe6
25.Bxh6 Rd8 26.Qf4 Qxf4 27.Bxf4 Nd5
28.Nxd5 Bxd5 29.Rd1 (4.282.793) 318

11/36 0:14 -6.72 22...Nh7 23.Rd1 Bxc3 24.bxc3 Nf6
25.Qxd6 Qxd6 26.Rxd6 Ke7 27.Rd3 Ra8
28.Nb3 e5 29.Bxh6 (4.801.210) 319

12/36 0:19 -6.94 22...Nh7 23.Nb3 Bxc3 24.bxc3 O-O
25.Rd1 Rc8 26.c4 Nf6 27.Qxd6 Nd5
28.Qxc7 Nc3+ 29.Kc1 Rxc7 (6.246.847) 323

12/36 0:22 -6.92 22...d5 23.exd5 O-O 24.dxe6 Bxe6
25.Bxh6 Rd8 26.Qf4 Qxf4 27.Bxf4 Bf8
28.Ra4 Bc5 29.Nb3 Bxb3 30.cxb3 (7.113.535) 322

12/37 0:24 -6.72 22...Ng8 23.Rd1 Bxc3 24.bxc3 Nf6
25.Qxd6 Qxd6 26.Rxd6 Ke7 27.Rd3 Ra8
28.Nb3 e5 29.Bxh6 (7.907.817) 324

13/37 0:33 -7.13 22...Ng8 23.Rd1 Ne7 24.Bd4 e5 25.Bb6 Qc6
26.Nb3 Be6 27.Qxd6 Qxd6 28.Rxd6 O-O
29.Ra4 Bxb3 30.cxb3 (11.013.579) 330

13/37 0:45 -7.04 22...Nh7 23.Nb3 Bxc3 24.bxc3 O-O
25.Rd1 Rc8 26.c4 Nf6 27.Qxd6 Nd5
28.Qxc7 Nc3+ 29.Kc1 Rxc7 30.Rb8+ Kg7 (15.068.378) 330

13/37 0:52 -6.77 22...Bf8 23.Nb5 Bxb5 24.Rxb5 h5
25.Qd4 Be7 26.g5 e5 27.Qb4 Nd7
28.Rb7 Qc6 29.Nb3 (17.247.999) 330

14/42 1:08 -7.00 22...Bf8 23.Nb5 Bxb5 24.Rxb5 Be7
25.Ra5 Qc8 26.Nb3 h5 27.g5 Nd7
28.Qe2 Ne5 29.Rd1 Qc4 30.Ra8+ Bd8 (22.790.676) 332

14/45 2:09 -6.88 22...Kf8 23.Bb6 Qc6 24.Nb3 Ne8
25.Bd4 e5 26.Be3 h5 27.Na5 Qa6
28.Rb7 Nf6 29.Rb8+ Bc8 (43.228.521) 334

15/45 2:34 -7.03 22...Kf8 23.Bb6 Qc6 24.Nb3 Ne8
25.Bd4 h5 26.g5 Bxd4 27.Nxd4 Qa6
28.f4 Ke7 29.f5 gxf5 30.exf5 (51.848.694) 336

16/45 5:05 -6.96 22...Kf8 23.Bb6 Qc6 24.Nb3 Ne8
25.Bd4 h5 26.g5 e5 27.Bf2 Kg8 28.Nd5 Kh7
29.Na5 Qa6 30.Rb7 (103.759.004) 339

17/52 8:17 -7.03 22...Kf8 23.Bb6 Qc6 24.Nb3 Ne8
25.Bd4 h5 26.g5 e5 27.Bf2 Be6 28.Nd5 Bxd5
29.exd5 Qa6 30.Bb6 Kg8 31.Re1 (169.153.878) 339

22... Kf8 chosen later, 22... Ng8 or 22... Nh7, or 22... d5, 22... O-O chosen instead of 22..Bf8 23.Nb5

4k2r/2qb1pb1/3ppnpp/8/1R2P1PP/2N1BP2/1PPQ4/NK5R b k -

Engine: Fruit 05/11/03 (128 MB)
by Fabien Letouzey

1/08 0:00 -5.66 22...Bc6 (138)

2/08 0:00 -5.82 22...Bc6 23.Nb3 (254)

2/08 0:00 -5.81 22...O-O 23.Bxh6 (362)

2/08 0:00 -5.38 22...h5 23.Bf4 (570)

3/11 0:00 -6.00 22...h5 23.g5 Nh7 (1.216)

4/11 0:00 -6.06 22...h5 23.g5 Nh7 24.Nb3 Bxc3
25.Qxc3 Qxc3 26.bxc3 (2.726)

4/11 0:00 -6.01 22...e5 23.Bb6 Qb8 24.Nb3 (4.180)

5/13 0:00 -6.41 22...e5 23.Nb3 Bc6 24.Nb5 Bxb5
25.Rxb5 (9.873)

5/13 0:00 -6.10 22...h5 23.g5 Nh7 24.Bd4 e5 (11.333)

6/13 0:00 -6.05 22...h5 23.g5 Ng8 24.Bd4 e5 25.Nd5 (18.292)

7/15 0:00 -6.47 22...h5 23.g5 Nh7 24.Bd4 O-O 25.Bxg7 Kxg7
26.Nb3 (34.274)

7/20 0:00 -6.31 22...Bc6 23.Nb3 Nd7 24.Nd4 Ke7
25.Ncb5 Bxb5 26.Rxb5 (56.077)

8/20 0:00 -6.24 22...Bc6 23.Nb3 Nd7 24.Nd4 h5
25.Nxc6 Qxc6 26.Bd4 Bxd4 27.Rxd4 (122.581)

8/23 0:00 -6.18 22...Bf8 23.Nb3 d5 24.Rd4 dxe4
25.Nxe4 Nxe4 26.Rxe4 Bc6 27.Qc3 (177.548)

9/23 0:00 -6.36 22...Bf8 23.Nb3 d5 24.Rd4 Qg3 25.Qe2 dxe4
26.Nxe4 Nxe4 27.Rxe4 Be7 (240.310)

10/26 0:01 -6.62 22...Bf8 23.Nb5 Bxb5 24.Rxb5 d5
25.Ra5 Qd8 26.Rd1 Be7 27.Qc3 Bd6
28.exd5 exd5 (613.691) 320

10/26 0:02 -6.52 22...h5 23.g5 Ng8 24.Bd4 e5 25.Nd5 Qc6
26.Rb8+ Bc8 27.Nb6 Ne7 28.Nxc8 Nxc8 (930.001) 315

10/26 0:03 -6.46 22...O-O 23.Bf4 Ra8 24.Bxd6 Qc6
25.Nb3 h5 26.Rg1 hxg4 27.fxg4 Nd5
28.exd5 Qxd6 29.dxe6 Qxe6 (1.216.435) 316

11/28 0:07 -7.38 22...O-O 23.Bxh6 Ra8 24.Bxg7 Kxg7
25.Nb3 Kf8 26.Qh6+ Ke7 27.Rd1 Bc6
28.Qf4 Bb7 (2.355.815) 322

11/29 0:07 -6.74 22...h5 23.g5 Ng8 24.Bd4 Bxd4
25.Rxd4 Ne7 26.Rd1 Nc8 27.Nb3 Bc6
28.Qf4 (2.553.961) 322

11/29 0:09 -6.51 22...Bf8 23.Nb5 Bxb5 24.Rxb5 h5
25.Qb4 Nd7 26.Nb3 Qc8 27.gxh5 d5
28.Qa4 (2.907.817) 322

12/29 0:12 -6.70 22...Bf8 23.Nb5 Bxb5 24.Rxb5 h5
25.Qd4 Be7 26.g5 e5 27.Qa4 Nd7
28.Nb3 O-O (4.018.980) 324

13/37 0:32 -6.83 22...Bf8 23.Rb6 e5 24.Nb3 Be7 25.Ra6 Qb8
26.Bxh6 Qb7 27.Ra5 Bd8 28.Ra3 Bc7 (10.496.349) 327

14/41 1:17 -6.72 22...Bf8 23.Rb6 Be7 24.Nb3 Bc6
25.Ra6 Qc8 26.Ra7 h5 27.g5 Nd7
28.Na5 Ba8 29.Nc4 (25.487.461) 329

15/41 1:53 -6.91 22...Bf8 23.Rb6 Be7 24.Nb3 Bc6
25.Ra6 Qc8 26.Ra7 h5 27.g5 Nd7
28.Rd1 Rg8 29.Bd4 Qb8 (37.760.762) 331

15/46 3:17 -6.81 22...Kf8 23.Nb3 Ne8 24.Qd3 h5 25.g5 Bxc3
26.Rc4 Bb5 27.Rxc7 Bxd3 28.Rxc3 Ba6
29.Rc6 Bb5 (66.537.362) 336

16/46 4:46 -6.94 22...Kf8 23.Nb3 Ne8 24.Nd4 h5 25.g5 Kg8
26.Rd1 Kh7 27.Rb3 Qa5 28.Rb7 Qd8
29.Ra7 Nc7 30.Qh2 (96.581.040) 336

17/46 9:59 -6.98 22...Kf8 23.Nb3 Kg8 24.Rd1 Bf8
25.Nb5 Bxb5 26.Rxb5 d5 27.exd5 Nxd5
28.h5 gxh5 29.gxh5 Nxe3 30.Qxe3 Be7
31.f4 Kh7 32.Qe4+ Kg7 33.Rg1+ Kf8 (201.566.940) 335


For the default settings the analysis is unchanged. The output from Gambit 005 or any other changed settings is reproducible. There is no way of finding out what settings are used in tournament mode, unless maybe something can be found in ucidebug mode, but the output at least is the same everytime. But any parameter testing this way is not possible, at least not in Shredder, but maybe some of the other GUIs fare a little better?

It could still be a Shredder GUI problem alone, but then a very strange one and it would be very strange if nobody noticed anything like this before. That seems not possible to me.

Regards, Eelco

[mclane]

use arena beta4

[Eelco de Groot]

use arena beta4

Thanks Thorsten, sure I could do that, but I must say I am not very happy with a UCI-engine that does not work well in the Shredder UCI interface, even if it is not the latest version of Shredder GUI. For me that is a standard for UCI-engines. And it is clear that it is a traceable bug. I sent a mail to Shredder support, I hope that Stefan Meyer-Kahlen will find some time to look at it if there can't be found any other solution. I have not yet sent a mail to Ryan.

I did download Arena 1.99 Beta 4 and the program started up as it should, -I am missing any installation instructions folks!-, but I would like to limit the number of engine interfaces a bit, unless I have no choice for testing Fruit. Arena seems a nice program though, I have not studied any of the set-up procedures yet, but e.g. as far as I know it is the only interface that would allow time handicap matches, I think that would be interesting for instance to simulate extremely strong opponents, testing when engines differ a lot in strength, giving the opponent a lot of time if you want to "crash-test" your own booklines etc.

Here are my Gambit 005 settings, I have not yet tested them for above-mentioned reasons, so it is only experimental but I think it works okay in analysis mode, although of course in analysis also very limited testing. The games played so far in the Shredder interface were clearly not using anything correctly set, as I posted above.

Some of the options in the settings are suggestions from Eduard Nemeth, some are analogous to Toga Blueberry settings, but maybe better things can be found because there are evidently a lot of possibilities for experiments.


Fruit 05/11/03 Gambit 005

Hash = 256
NalimovPath = C:\Nalimov
NalimovCache = 4
BookFile = book.bin
Time Allocation = 100
EGTB = true
EGTB Depth = 4 [Default = 8]
NullMove Pruning = Always
NullMove Reduction = 3
Verification Search = Always
Verification Reduction = 6 [Default = 5]
History Pruning 1 = true
History Depth 1 = 1
History PV Node 1 = false
History Move Skip 1 = 3
History Treshold 1 = 40
History Reduction 1 = 2
History Leaf Pruning 1 = true
History Pruning 2 = true
History Depth 2 = 1
History PV Node 2 = false
History Move Skip 2 = 5
History Treshold 2 = 70
History Reduction 2 = 1
History Leaf Pruning 2 = true
History ReSearch = true
Extended Futility Pruning = true
Extended Futility Margin = 300
Material = 102 [Default =100]
Piece Activity = 100
King safety = 125 [Default = 100]
Pawn Structure = 102 [Default = 100]
Passed Pawns = 102 [Default =100]
Pawn = 100
Knight = 100
Bishop = 100
Rook = 100
Queen = 100
Bishop Pair = 100
Pawn Duo = true [Default = false]
Pawn Duo Weight = 4
Pawn Duo Weight (Endgame) = 1 [Default = 0]
Pawn Duo File a+b = 1
Pawn Duo File b+c = 2
Pawn Duo File c+d = 3
Pawn Duo File d+e = 4
Pawn Duo Rank 2 = 0
Pawn Duo Rank 3 = 0
Pawn Duo Rank 4 = 4
Pawn Duo Rank 5 = 5
Pawn Duo Rank 6 = 6
Pawn Duo Rank 7 = 7
Contempt Factor = 0 [Actually it is set at 0.05 in ChessPartner but I don't think this is an allowed value. Shredder allows integers 1-5. For all parameters: it's untested in games]