Last Update: 23.05.2022
Hello Chess Friends,
As usual, I am pleased to announced that,This time, I realized to run several but various testings...As NNUE and NON-NNUE, Ponder ON / Ponder OFF As alternative, exc. Cutechess, I used Fritz GUI too
1st of all,For a better conclusion, more games (per player) required..E.g 1000 games sounds not so bad...but that's all from MeAnyhow, I hope all the current new results will be useful...
To be more clear,I was wondering what will be the influence in case of testingsUnder different conditions ..actually no BIG surprise but anyhow,As we see, Chucaro in most testings produced very close points,Sure, we can't expect exactly same points and I think it's normal
On other hand,When I say in most tests close points, I mean about performance..But when NNUE: the draw ratio is too high, especially by SF-PBRegarding in case of test as NON -NNUE: minus plus 15% less..
By the way, let Me say please a few words about Ponder OFF/ONThe results indicate that there is no any BIG influence in case ofWhen Ponder is enabled or disabled e.g almost identical in points..But no doubt that if Ponder ON: better/stronger games for sure !)Proved, guaranteed, especially if testings are NON-NNUE: Results
Meanwhile, under Fritz GUI, as NON-NNUE (via Cfish 030820),The draw ratio is recorded 4-5% less than Cutechess...interesting)
And why I used Chucaro 3.6a (by Angel Morano) for all testings,The reason, Chucaro is a Top book, quite new vers. in Bin formatOtherwise, if CTG, then Book Learning may affect all the resultsPlus in this way, I tested it under Cutechess and Fritz GUI as well
Ok..here are all the new played results: (420 games per test):
1st Test - Official Results of SCCT CS Book / NNUE I
KomodoDragon 2.6 : 212.5 POINTS / 91% DRAWS
Note: 1 Core, Ponder OFF, 1m+1s, 64 MB Hash, Cutechess GUIChucaro book played by Komodo, the opponents by Cfish 200621 ----------------------------------------------------------------------------------------
2nd Test: Target to see the influence via other NNUE engine
SF PB 190422 : 217.0 POINTS / 95% DRAWS
Note: 1 Core, Ponder OFF, 1m+1s, 64 MB Hash, Cutechess GUIAll books played by SF PB 190422 engine (via default Evalfile)----------------------------------------------------------------------------------------
3rd Test - Target to see influence as NON-NNUE under Fritz:
Cfish 030820 C40 : 215.5 POINTS / 77% DRAWS
Note: 2 Core/Game, Ponder ON, 2m+1s, 128 MB Hash, Fritz GUIAll books played by Cfish 030820 (including the below testings too)----------------------------------------------------------------------------------------
4th Test - Target to see the influence NON-NNUE under Cutechess
Cfish 030820 C40 : 222.5 POINTS / 81% DRAWS Note: 2 Core/Game, Ponder ON, 2m+1s, 128 MB Hash, Cutechess----------------------------------------------------------------------------------------
5th Test - The Target to see the influence as Ponder OFF:
Cfish 030820 C40 : 222.0 POINTS / 82% DRAWS
Note: 1 Core, Ponder OFF, 2m+1s, 128 MB Hash, Cutechess----------------------------------------------------------------------------------------
More Details:All gauntlet testings played via same BIN book: Chucaro 3.6aChucaro's opponents belongs to Championship and as usual:For each opening book player is installed separate exe engineAnd no any game lost on time, it seems the conditions are stableAll the played games (2100 games in PGN) available: Download
As Final Conclusion,Any book's performance depends on our used conditions, e.g:Time Control, MP or 1 Core, NN or NON etc. even any EngineOr GUI can make some influences over the played results...
Btw, without to not mention this I can't too, e.g in case of test asUnexpected, disappointed results, often the testers are blamed..Sure...no much free time to count all, but just a few examples:Some call it as tester/tour errors...)) some not happy with size limitsOr with the amount of played games or with the used/tested engines Or with my used time controls,..btw, I know very well how to increaseThe quality, but just in case..this time I'm afraid that to see a win: It will be something like as looking for a needle in a haystack ))Plus I will be not able to run and test so many opening books orSome are not happy why I test books from the past to the present..Btw, what about if I will start to look for holes over their work..??))You may know...I have some experience in Computer chess too )In other words, over the past long years, I've seen a lot of critics...Some of these critics can be counted as useful..but mostly are not...And sorry that I can not satisfy all..really...it's just impossible...!Yes..my work is not perfect...but I am doing my best and if there isSomething wrong..just let Me know, but with better ideas please..Sure..this is also true that we have many book makers who are realExperts and as usual, they are too busy with improving their books..Because success comes through hard work and who never give up!
As final words, SCCT is organized the 1st ever Book Championships!And if I caused to appear many Book Tours by others, how happy I'm
Ok...that's all for now....and thanks for your interest...
Best,Sedat
Hello Chess Friends,
As usual, I am pleased to announced that,This time, I realized to run several but various testings...As NNUE and NON-NNUE, Ponder ON / Ponder OFF As alternative, exc. Cutechess, I used Fritz GUI too
1st of all,For a better conclusion, more games (per player) required..E.g 1000 games sounds not so bad...but that's all from MeAnyhow, I hope all the current new results will be useful...
To be more clear,I was wondering what will be the influence in case of testingsUnder different conditions ..actually no BIG surprise but anyhow,As we see, Chucaro in most testings produced very close points,Sure, we can't expect exactly same points and I think it's normal
On other hand,When I say in most tests close points, I mean about performance..But when NNUE: the draw ratio is too high, especially by SF-PBRegarding in case of test as NON -NNUE: minus plus 15% less..
By the way, let Me say please a few words about Ponder OFF/ONThe results indicate that there is no any BIG influence in case ofWhen Ponder is enabled or disabled e.g almost identical in points..But no doubt that if Ponder ON: better/stronger games for sure !)Proved, guaranteed, especially if testings are NON-NNUE: Results
Meanwhile, under Fritz GUI, as NON-NNUE (via Cfish 030820),The draw ratio is recorded 4-5% less than Cutechess...interesting)
And why I used Chucaro 3.6a (by Angel Morano) for all testings,The reason, Chucaro is a Top book, quite new vers. in Bin formatOtherwise, if CTG, then Book Learning may affect all the resultsPlus in this way, I tested it under Cutechess and Fritz GUI as well
Ok..here are all the new played results: (420 games per test):
1st Test - Official Results of SCCT CS Book / NNUE I
KomodoDragon 2.6 : 212.5 POINTS / 91% DRAWS
Note: 1 Core, Ponder OFF, 1m+1s, 64 MB Hash, Cutechess GUIChucaro book played by Komodo, the opponents by Cfish 200621 ----------------------------------------------------------------------------------------
2nd Test: Target to see the influence via other NNUE engine
SF PB 190422 : 217.0 POINTS / 95% DRAWS
Note: 1 Core, Ponder OFF, 1m+1s, 64 MB Hash, Cutechess GUIAll books played by SF PB 190422 engine (via default Evalfile)----------------------------------------------------------------------------------------
3rd Test - Target to see influence as NON-NNUE under Fritz:
Cfish 030820 C40 : 215.5 POINTS / 77% DRAWS
Note: 2 Core/Game, Ponder ON, 2m+1s, 128 MB Hash, Fritz GUIAll books played by Cfish 030820 (including the below testings too)----------------------------------------------------------------------------------------
4th Test - Target to see the influence NON-NNUE under Cutechess
Cfish 030820 C40 : 222.5 POINTS / 81% DRAWS Note: 2 Core/Game, Ponder ON, 2m+1s, 128 MB Hash, Cutechess----------------------------------------------------------------------------------------
5th Test - The Target to see the influence as Ponder OFF:
Cfish 030820 C40 : 222.0 POINTS / 82% DRAWS
Note: 1 Core, Ponder OFF, 2m+1s, 128 MB Hash, Cutechess----------------------------------------------------------------------------------------
More Details:All gauntlet testings played via same BIN book: Chucaro 3.6aChucaro's opponents belongs to Championship and as usual:For each opening book player is installed separate exe engineAnd no any game lost on time, it seems the conditions are stableAll the played games (2100 games in PGN) available: Download
As Final Conclusion,Any book's performance depends on our used conditions, e.g:Time Control, MP or 1 Core, NN or NON etc. even any EngineOr GUI can make some influences over the played results...
Btw, without to not mention this I can't too, e.g in case of test asUnexpected, disappointed results, often the testers are blamed..Sure...no much free time to count all, but just a few examples:Some call it as tester/tour errors...)) some not happy with size limitsOr with the amount of played games or with the used/tested engines Or with my used time controls,..btw, I know very well how to increaseThe quality, but just in case..this time I'm afraid that to see a win: It will be something like as looking for a needle in a haystack ))Plus I will be not able to run and test so many opening books orSome are not happy why I test books from the past to the present..Btw, what about if I will start to look for holes over their work..??))You may know...I have some experience in Computer chess too )In other words, over the past long years, I've seen a lot of critics...Some of these critics can be counted as useful..but mostly are not...And sorry that I can not satisfy all..really...it's just impossible...!Yes..my work is not perfect...but I am doing my best and if there isSomething wrong..just let Me know, but with better ideas please..Sure..this is also true that we have many book makers who are realExperts and as usual, they are too busy with improving their books..Because success comes through hard work and who never give up!
As final words, SCCT is organized the 1st ever Book Championships!And if I caused to appear many Book Tours by others, how happy I'm
Ok...that's all for now....and thanks for your interest...
Best,Sedat