Latest Update: 15.12.2020 Strength Testings (played mainly for XLVII).: Details
Dear Chess Friends,
Just I'd like to inform you that,XLVII tour is already ended! Congrats and many thanks to all!
Frankly, I've lost my interest to work over current book tourneyThe reason, due to high-draw percentage..really too much boringSure I am not against draws, but when they are too much...sorry...Even 1 core did not make BIG influence over producing less drawsPlus, I can't call it as real book tournament, in case of NNUE engineThe influence is maybe minus plus 70-80% EvalFile / 20-30% BookE.g, if we check No Book's performance...it is clear that the EvalFilePlays more important role, regarding the influence over many books...But in case of test as NON-NNUE then the game is totally different!)E.g, No Book's performance can be compared: 45 36 34 30 23 19 etc.Note: No Book means the engine is played without using any book
As I mentioned before, I believe in that (anyone who missed):As NNUE, it's expecting to see stronger, better quality games!However, via NNUE, it does not mean as 'best' book testings...In other words, we should not mix apples with oranges!) I mean:Via NNUE is not ideal way for checking the opening books lines!Because, many many of the winning lines are become as drawishThat means also many of the loosing lines started to be drawish tooE.g, in most cases: with very critical (weak) lines, we can see 'wins'Sure, there are some other ways to reduce the draw percentage e.gIn case of playing with ultra fast time controls etc...or playingVs weaker opponents etc..but for nowadays I have no such plans...So, I can not call it as much logical, and it seems the EvalFileHelps to NNUE engine to prefer stronger moves, but in Book tours:I think that the opening books should not be check with extra file!Let's allow pure playing, otherwise it's not so much original too!Btw, just one example, which perfectly suits over this issue:-Are the students allowed to look at any file (in the Exam?)
Sure, I am not against NNUE book testings...but before,The EvalFiles should start to produce less draws...otherwise:No much fun, at least for my side. thanks for your understanding!
However, I have good news,Soon I hope to organize another new book tour...More info coming soon!
XLVII's Champion: Perfec book- My Congratulations to Cevdet Sarı!
# PLAYER RATING POINTS GAMES (%) 1 Perfect by Cevdet : 3715.4 265.0 500 53.0% 2 2600-Club : 3714.9 137.0 260 52.7% 3 Outskirts by Kerveros : 3713.6 136.5 260 52.5% 4 Lion 5 : 3711.9 235.5 450 52.3% 5 W&B 3456 : 3710.3 234.5 450 52.1% 6 Goi Out.081220 : 3709.6 135.0 260 51.9% 7 M9.7 : 3708.2 233.0 450 51.8% 8 RadaR 011220 : 3707.9 254.5 490 51.9% 9 ChazzChess : 3706.7 232.0 450 51.6% 10 Tokugawa-Beta : 3706.7 232.0 450 51.6% 11 Snow-Ball : 3705.7 133.5 260 51.3% 12 Fidelio : 3704.4 230.5 450 51.2% 13 OLiprog : 3703.1 132.5 260 51.0% 14 Ceberelıum 3M : 3700.6 260.0 510 51.0% 15 No Book : 3700.0 227.5 450 50.6% 16 Fres : 3698.7 207.5 410 50.6% 17 DON Lite 2 : 3698.2 257.5 510 50.5% 18 QuarterPounder : 3697.6 257.0 510 50.4% 19 Goi Out.221120 : 3697.3 180.0 360 50.0% 20 Rufus 231219 : 3697.1 256.5 510 50.3% 21 Fairey Battle : 3689.4 176.0 360 48.9% 22 NebulaBlue76x : 3681.8 168.0 350 48.0% 23 Manifest 260920b : 3680.8 167.5 350 47.9% 24 LC-X 8.1a : 3679.8 190.0 400 47.5% 25 HIPPO v1.1 : 3677.0 122.5 260 47.1% 26 Cobra Kai II : 3670.5 120.0 260 46.2% 27 Buffalo : 3670.0 162.0 350 46.3% 28 MiniPatrolT : 3667.4 182.5 400 45.6% 29 LC-X 8.1f : 3666.5 118.5 260 45.6% 30 MiniPatrol 2a : 3661.3 116.5 260 44.8% 31 Cobra Kai : 3658.2 89.0 200 44.5%
More Details: • All books are played by same chess engine: CF EXT 161120 NNUE • CF EXT engine is used to play via EvalFile: nn-cb26f10b1fd9.nnue • Maximum uncompressed size limit CTG: 1200 MB / BIN: 250 MB • Chessbase (ctg) main book settings: Learning ON, Optimized, 0/100 • Exception, some ctg ones are payed by Author's recommendations • Polyglot (bin) formats are tested with best book move as disabled • Current overall draw percentage(based on 5850 games) is high 93% • In case of checking the Top books draws (played each other) is 97% • It is clear that mostly defeats come from the critical (weaker) books • All the played tournament games are available (in PGN): XLVII
Dear Chess Friends,
Just I'd like to inform you that,XLVII tour is already ended! Congrats and many thanks to all!
Frankly, I've lost my interest to work over current book tourneyThe reason, due to high-draw percentage..really too much boringSure I am not against draws, but when they are too much...sorry...Even 1 core did not make BIG influence over producing less drawsPlus, I can't call it as real book tournament, in case of NNUE engineThe influence is maybe minus plus 70-80% EvalFile / 20-30% BookE.g, if we check No Book's performance...it is clear that the EvalFilePlays more important role, regarding the influence over many books...But in case of test as NON-NNUE then the game is totally different!)E.g, No Book's performance can be compared: 45 36 34 30 23 19 etc.Note: No Book means the engine is played without using any book
As I mentioned before, I believe in that (anyone who missed):As NNUE, it's expecting to see stronger, better quality games!However, via NNUE, it does not mean as 'best' book testings...In other words, we should not mix apples with oranges!) I mean:Via NNUE is not ideal way for checking the opening books lines!Because, many many of the winning lines are become as drawishThat means also many of the loosing lines started to be drawish tooE.g, in most cases: with very critical (weak) lines, we can see 'wins'Sure, there are some other ways to reduce the draw percentage e.gIn case of playing with ultra fast time controls etc...or playingVs weaker opponents etc..but for nowadays I have no such plans...So, I can not call it as much logical, and it seems the EvalFileHelps to NNUE engine to prefer stronger moves, but in Book tours:I think that the opening books should not be check with extra file!Let's allow pure playing, otherwise it's not so much original too!Btw, just one example, which perfectly suits over this issue:-Are the students allowed to look at any file (in the Exam?)
Sure, I am not against NNUE book testings...but before,The EvalFiles should start to produce less draws...otherwise:No much fun, at least for my side. thanks for your understanding!
However, I have good news,Soon I hope to organize another new book tour...More info coming soon!
XLVII's Champion: Perfec book- My Congratulations to Cevdet Sarı!
# PLAYER RATING POINTS GAMES (%) 1 Perfect by Cevdet : 3715.4 265.0 500 53.0% 2 2600-Club : 3714.9 137.0 260 52.7% 3 Outskirts by Kerveros : 3713.6 136.5 260 52.5% 4 Lion 5 : 3711.9 235.5 450 52.3% 5 W&B 3456 : 3710.3 234.5 450 52.1% 6 Goi Out.081220 : 3709.6 135.0 260 51.9% 7 M9.7 : 3708.2 233.0 450 51.8% 8 RadaR 011220 : 3707.9 254.5 490 51.9% 9 ChazzChess : 3706.7 232.0 450 51.6% 10 Tokugawa-Beta : 3706.7 232.0 450 51.6% 11 Snow-Ball : 3705.7 133.5 260 51.3% 12 Fidelio : 3704.4 230.5 450 51.2% 13 OLiprog : 3703.1 132.5 260 51.0% 14 Ceberelıum 3M : 3700.6 260.0 510 51.0% 15 No Book : 3700.0 227.5 450 50.6% 16 Fres : 3698.7 207.5 410 50.6% 17 DON Lite 2 : 3698.2 257.5 510 50.5% 18 QuarterPounder : 3697.6 257.0 510 50.4% 19 Goi Out.221120 : 3697.3 180.0 360 50.0% 20 Rufus 231219 : 3697.1 256.5 510 50.3% 21 Fairey Battle : 3689.4 176.0 360 48.9% 22 NebulaBlue76x : 3681.8 168.0 350 48.0% 23 Manifest 260920b : 3680.8 167.5 350 47.9% 24 LC-X 8.1a : 3679.8 190.0 400 47.5% 25 HIPPO v1.1 : 3677.0 122.5 260 47.1% 26 Cobra Kai II : 3670.5 120.0 260 46.2% 27 Buffalo : 3670.0 162.0 350 46.3% 28 MiniPatrolT : 3667.4 182.5 400 45.6% 29 LC-X 8.1f : 3666.5 118.5 260 45.6% 30 MiniPatrol 2a : 3661.3 116.5 260 44.8% 31 Cobra Kai : 3658.2 89.0 200 44.5%
• Calculated by Ordo / No Book is fixed to 3700 Elo / Blue Name = Last Update
More Details: • All books are played by same chess engine: CF EXT 161120 NNUE • CF EXT engine is used to play via EvalFile: nn-cb26f10b1fd9.nnue • Maximum uncompressed size limit CTG: 1200 MB / BIN: 250 MB • Chessbase (ctg) main book settings: Learning ON, Optimized, 0/100 • Exception, some ctg ones are payed by Author's recommendations • Polyglot (bin) formats are tested with best book move as disabled • Current overall draw percentage(based on 5850 games) is high 93% • In case of checking the Top books draws (played each other) is 97% • It is clear that mostly defeats come from the critical (weaker) books • All the played tournament games are available (in PGN): XLVII
46 45 44 43 42 41 40 39 38 37 36 35 34 33 32 31 30 29 28 27 26 25 24 23