Some sloppy results

[Tony Thomas]

Sloppy is one of the strongest open source engine that emerged recently. Despite all the clone brujaaha, it doesnt play much like Fruit or any of its derivatives. If it actually played like fruit, it should be at least as strong. Currently Sloppy is not a top 30 engine, it is much weaker than the likes of Movei and Bugchess. To get an idea of Sloppy's strength I post 9 other engines that are in the same league as sloppy, note that some of them are outdated and plays at a higher division.

Code: Select all

26 Zarkov 4.89                    : 2559   43  43   195    51.3 %   2550   22.6 %
 27 Ufim 8.02                      : 2559   43  43   195    51.3 %   2550   21.5 %
 28 Tao 5.6 Blitz Specialist       : 2555   43  43   197    50.8 %   2550   21.3 %
 29 Crafty 21.4 MB                 : 2550   56  56   114    50.0 %   2550   24.6 %
 30 Colossus2006f                  : 2547   43  43   187    50.3 %   2545   24.6 %
 31 Sloppy-0.1.1 JA                : 2546   52  52   130    48.1 %   2560   25.4 %
 32 Amyan 1.597                    : 2543   42  42   199    49.2 %   2548   26.1 %
 33 Yace.099.56                    : 2540   44  44   195    48.5 %   2550   21.0 %
 34 Chiron v0.8.1b6                : 2533   54  54   121    48.8 %   2542   26.4 %
 35 Green Light Chess 3.01.2.2     : 2531   42  42   199    47.5 %   2549   23.6 %
 36 Baron 1.70                     : 2529   46  46   167    47.3 %   2547   24.0 %

Code: Select all

31 Sloppy-0.1.1 JA           : 2546  130 (+ 46,= 33,- 51), 48.1 %

Amyan 1.597                   :   4 (+  1,=  2,-  1), 50.0 %
Alaric 061011                 :   4 (+  1,=  2,-  1), 50.0 %
Arasan 9.4                    :   4 (+  4,=  0,-  0), 100.0 %
Comet_B68                     :   4 (+  3,=  0,-  1), 75.0 %
Green Light Chess 3.01.2.2    :   4 (+  1,=  1,-  2), 37.5 %
Patzer 3.80                   :   4 (+  2,=  1,-  1), 62.5 %
Nejmet_307                    :   4 (+  0,=  1,-  3), 12.5 %
Trace 1.36 NC3                :   4 (+  1,=  1,-  2), 37.5 %
Crafty Cito 2.0               :   3 (+  0,=  2,-  1), 33.3 %
Quark 2.35                    :   4 (+  3,=  1,-  0), 87.5 %
Zarkov 4.89                   :   4 (+  0,=  2,-  2), 25.0 %
E.T.Chess 18.10.05            :   4 (+  1,=  1,-  2), 37.5 %
AnMon 5.60                    :   4 (+  1,=  0,-  3), 25.0 %
Petir 3.99d                   :   4 (+  2,=  2,-  0), 75.0 %
Yace.099.56                   :   4 (+  3,=  0,-  1), 75.0 %
Pepit 1.59                    :   4 (+  2,=  0,-  2), 50.0 %
Zappa 1.1                     :   4 (+  3,=  1,-  0), 87.5 %
Ufim 8.02                     :   4 (+  3,=  0,-  1), 75.0 %
Frenzee 3.0                   :   4 (+  2,=  0,-  2), 50.0 %
Tao 5.6 Blitz Specialist      :   4 (+  0,=  1,-  3), 12.5 %
WildCat 6.0                   :   4 (+  1,=  1,-  2), 37.5 %
Pseudo07c                     :   4 (+  1,=  0,-  3), 25.0 %
Chiron v0.8.1b10.3            :   4 (+  1,=  3,-  0), 62.5 %
Sjeng 1.6                     :   4 (+  1,=  1,-  2), 37.5 %
Bright 0.1                    :   4 (+  0,=  3,-  1), 37.5 %
Crafty 21.5 MB                :   3 (+  0,=  2,-  1), 33.3 %
Booot 4.13.0                  :   4 (+  1,=  0,-  3), 25.0 %
Baron 1.8.1                   :   4 (+  2,=  1,-  1), 62.5 %
RomiChessDK5                  :   4 (+  3,=  0,-  1), 75.0 %
Movei00_8_403 10 10 10        :   4 (+  1,=  1,-  2), 37.5 %
Colossus2007d                 :   4 (+  1,=  2,-  1), 50.0 %
BugChess2_V1_5_2              :   4 (+  0,=  0,-  4),  0.0 %
Crafty 21.6 JA                :   4 (+  1,=  1,-  2), 37.5 %

[Dann Corbit]

Sloppy has an excellent bookbuilder, and it utilizes some statistics. If you build a good book for sloppy, it will add more than 100 Elo.

With the book that I built, the strength is identical to ProDeo.

[ilari]

Sloppy has an excellent bookbuilder, and it utilizes some statistics. If you build a good book for sloppy, it will add more than 100 Elo.

With the book that I built, the strength is identical to ProDeo.

I've now replaced the book on my site with a better one, built from this PGN file:
http://cvs.mandriva.com/cgi-bin/viewvc. ... gm2600.pgn

The next version of Sloppy will use a different book format, as you have suggested.

[ilari]

What time control did you use? Sloppy is currently very weak at "whole game in X minutes" time controls. I have that fixed now, but the fix will have to wait until the next release.

[Tony Thomas]

What time control did you use? Sloppy is currently very weak at "whole game in X minutes" time controls. I have that fixed now, but the fix will have to wait until the next release.

I always use the same time control, Im+1sec. I used the book that came with JA compile.

[ilari]

What time control did you use? Sloppy is currently very weak at "whole game in X minutes" time controls. I have that fixed now, but the fix will have to wait until the next release.

I always use the same time control, Im+1sec. I used the book that came with JA compile.

Yes, Sloppy seems to have big problems with that time control. And the book in the JA compile is probably the same one I made previously.

I just ran 100 (40 moves/2 min) games with the new book against BugChess: Sloppy-0.1.1 vs. BugChess2 V1.5.2: 55-25-20. So Sloppy won 65% of points. I always test with the 64-bit Linux version.

[Tony Thomas]

What time control did you use? Sloppy is currently very weak at "whole game in X minutes" time controls. I have that fixed now, but the fix will have to wait until the next release.

I always use the same time control, Im+1sec. I used the book that came with JA compile.

Yes, Sloppy seems to have big problems with that time control. And the book in the JA compile is probably the same one I made previously.

I just ran 100 (40 moves/2 min) games with the new book against BugChess: Sloppy-0.1.1 vs. BugChess2 V1.5.2: 55-25-20. So Sloppy won 65% of points. I always test with the 64-bit Linux version.

Next time you release a version, I will test it with the new book. However, I cant change my time control. Do you have any plans to fix incremental time controls? My que is as long as is, here is the beginning of the list.

Code: Select all

Prodeo 1.5
Naum (Tiny book)
Xpdnt
Alfil beta
firefily
Surprise
BBchess
Typhoon
Pupsi
Twisted Logic
Petir 4.999
Homer 2.01
Flux
Dirty
Rotor
Delphil
Mediocre

[ilari]

Next time you release a version, I will test it with the new book. However, I cant change my time control. Do you have any plans to fix incremental time controls?

I already have "sort of" fixed the time management. The fix will be in the next version of Sloppy. I say "sort of" because Sloppy will still perform better at longer time controls, but that has very little to do with time management.

[Uri Blass]

Sloppy has an excellent bookbuilder, and it utilizes some statistics. If you build a good book for sloppy, it will add more than 100 Elo.

With the book that I built, the strength is identical to ProDeo.

How many games did you use to get the estimate of more than
100 elo?

Uri

[Tony Thomas]

Next time you release a version, I will test it with the new book. However, I cant change my time control. Do you have any plans to fix incremental time controls?

I already have "sort of" fixed the time management. The fix will be in the next version of Sloppy. I say "sort of" because Sloppy will still perform better at longer time controls, but that has very little to do with time management.

Any improvement in strength wont hurt, and it would make the clone accusers even more angry.