Laskos wrote: ↑Fri Aug 14, 2020 10:36 am
-mwyoung- is known to not change his mind at any cost even in the front of overwhelming evidence that he might do much better. Leave him alone, I never follow his tests, he used to use some 36 threads on 16 core machine, I don't know what other silly things he is doing.
I do silly crazy things like. Test chess engines much longer then 1 miniute per game. And test chess engines using more then 1 thread. Then I get even crazier by posting my settings and live streaming the engine test.
The only thing I noticed with your "tests" is that you get higher draw rate than a typical correspondence chess match of today.
That really makes them super uninteresting for anything.
Maybe it suggest that typical correspondence chess players are weaker than the top engines because they do not use a lot of hardware time or do not use the right software.
I doubt if it is possible to beat stockfish with no book in a correspondence game
I would like to know how much do you get with 10:1 time advantage with a good hardware
Let say stockfish with 10 minutes per move against stockfish with 1 minutes per move ponder off with a good hardware and if the 10 minutes per move can win at least 10% of the games or cannot do it.
Laskos wrote: ↑Fri Aug 14, 2020 10:36 am
-mwyoung- is known to not change his mind at any cost even in the front of overwhelming evidence that he might do much better. Leave him alone, I never follow his tests, he used to use some 36 threads on 16 core machine, I don't know what other silly things he is doing.
I do silly crazy things like. Test chess engines much longer then 1 miniute per game. And test chess engines using more then 1 thread. Then I get even crazier by posting my settings and live streaming the engine test.
The only thing I noticed with your "tests" is that you get higher draw rate than a typical correspondence chess match of today.
That really makes them super uninteresting for anything.
Maybe it suggest that typical correspondence chess players are weaker than the top engines because they do not use a lot of hardware time or do not use the right software.
I doubt if it is possible to beat stockfish with no book in a correspondence game
I would like to know how much do you get with 10:1 time advantage with a good hardware
Let say stockfish with 10 minutes per move against stockfish with 1 minutes per move ponder off with a good hardware and if the 10 minutes per move can win at least 10% of the games or cannot do it.
You mean from same standard opening position all the games? That's how correspondence chess works.
If openings are used with engines, it depends much on openings. There seem to be two main tendencies: using very good but drawish openings and using not so good but lower draw rate openings played side and reverse. To have many good openings, one has to use longer lines than in case of using many not so good openings. That results in very inefficient testing from "good openings" at anything longer than ultra-fast and fast TC.
By the way, rating groups like CCRL, CEGT, FGRL, SPPC etc. should standardize their opening repertoire, as the results depend much on it. I think SPPC is going at length optimizing its openings for efficiency. I think 2moves_v1 2-mover opening suite of Stockfish testing framework can safely be taken by all testing groups as standard, for efficiency and uniformity. It can too become drawish, especially in self-play and a bit longer TC, and then unbalanced openings should be introduced.
Laskos wrote: ↑Fri Aug 14, 2020 10:36 am
-mwyoung- is known to not change his mind at any cost even in the front of overwhelming evidence that he might do much better. Leave him alone, I never follow his tests, he used to use some 36 threads on 16 core machine, I don't know what other silly things he is doing.
I do silly crazy things like. Test chess engines much longer then 1 miniute per game. And test chess engines using more then 1 thread. Then I get even crazier by posting my settings and live streaming the engine test.
The only thing I noticed with your "tests" is that you get higher draw rate than a typical correspondence chess match of today.
That really makes them super uninteresting for anything.
Maybe it suggest that typical correspondence chess players are weaker than the top engines because they do not use a lot of hardware time or do not use the right software.
I doubt if it is possible to beat stockfish with no book in a correspondence game
I would like to know how much do you get with 10:1 time advantage with a good hardware
Let say stockfish with 10 minutes per move against stockfish with 1 minutes per move ponder off with a good hardware and if the 10 minutes per move can win at least 10% of the games or cannot do it.
You mean from same standard opening position all the games? That's how correspondence chess works.
If openings are used with engines, it depends much on openings. There seem to be two main tendencies: using very good but drawish openings and using not so good but lower draw rate openings played side and reverse. To have many good openings, one has to use longer lines than in case of using many not so good openings. That results in very inefficient testing from "good openings" at anything longer than ultra-fast and fast TC.
By the way, rating groups like CCRL, CEGT, FGRL, SPPC etc. should standardize their opening repertoire, as the results depend much on it. I think SPPC is going at length optimizing its openings for efficiency. I think 2moves_v1 2-mover opening suite of Stockfish testing framework can safely be taken by all testing groups as standard, for efficiency and uniformity. It can too become drawish, especially in self-play and a bit longer TC, and then unbalanced openings should be introduced.
I mean simply not use opening book and let the engines to choose the moves.
Stockfish is not deterministic with many cores so you can expect different games.
If you use some opening book that is the same for both engines then you need first to prove that this book is better than the empty book at long time control based on games.
Jouni wrote: ↑Wed Aug 12, 2020 9:36 pm
Yes SF NNUE is equal to quadruple your CPU cores for free. Incredible .
I actually got a result that SFNNUE (a couple days ago) on one thread beat Stockfish 11 on seven threads, at 2' + 1", by 90 to 80! So you may be understating it!
"The worst thing that can happen to a forum is a running wild attacking moderator(HGM) who is not corrected by the community." - Ed Schröder
But my words like silent raindrops fell. And echoed in the wells of silence.
Jouni wrote: ↑Fri Aug 14, 2020 1:45 pm
SF NNUE is beating now Lc0 badly +53 ELO. No need to use 90/176 cores like CCC or TCEC. All next finals SF vs SF NNUE!?
Stockfish+NNUE 170820 (Sergio 2257) Vs. Lc0 26.1 (J92-70) (32 thread test) (TC=15m+15s)
Hardware 2950x, RTX 2080 ti
Ponder off.
TC=15m+15s
32 threads.
8 Gb hash.
6 man TB.
Opening book Fritz 17 to 6 moves repeated for white and black.
Default settings.
"The worst thing that can happen to a forum is a running wild attacking moderator(HGM) who is not corrected by the community." - Ed Schröder
But my words like silent raindrops fell. And echoed in the wells of silence.
Jouni wrote: ↑Fri Aug 14, 2020 1:45 pm
SF NNUE is beating now Lc0 badly +53 ELO. No need to use 90/176 cores like CCC or TCEC. All next finals SF vs SF NNUE!?
Stockfish+NNUE 170820 (Sergio 2257) Vs. Lc0 26.1 (J92-70) (32 thread test) (TC=15m+15s)
Hardware 2950x, RTX 2080 ti
Ponder off.
TC=15m+15s
32 threads.
8 Gb hash.
6 man TB.
Opening book Fritz 17 to 6 moves repeated for white and black.
Default settings.
"The worst thing that can happen to a forum is a running wild attacking moderator(HGM) who is not corrected by the community." - Ed Schröder
But my words like silent raindrops fell. And echoed in the wells of silence.
Jouni wrote: ↑Fri Aug 14, 2020 1:45 pm
SF NNUE is beating now Lc0 badly +53 ELO. No need to use 90/176 cores like CCC or TCEC. All next finals SF vs SF NNUE!?
Stockfish+NNUE 170820 (Sergio 2257) Vs. Lc0 26.1 (J92-70) (32 thread test) (TC=15m+15s)
Hardware 2950x, RTX 2080 ti
Ponder off.
TC=15m+15s
32 threads.
8 Gb hash.
6 man TB.
Opening book Fritz 17 to 6 moves repeated for white and black.
Default settings.
"The worst thing that can happen to a forum is a running wild attacking moderator(HGM) who is not corrected by the community." - Ed Schröder
But my words like silent raindrops fell. And echoed in the wells of silence.
Jouni wrote: ↑Fri Aug 14, 2020 1:45 pm
SF NNUE is beating now Lc0 badly +53 ELO. No need to use 90/176 cores like CCC or TCEC. All next finals SF vs SF NNUE!?
Stockfish+NNUE 170820 (Sergio 2257) Vs. Lc0 26.1 (J92-70) (32 thread test) (TC=15m+15s)
Hardware 2950x, RTX 2080 ti
Ponder off.
TC=15m+15s
32 threads.
8 Gb hash.
6 man TB.
Opening book Fritz 17 to 6 moves repeated for white and black.
Default settings.
"The worst thing that can happen to a forum is a running wild attacking moderator(HGM) who is not corrected by the community." - Ed Schröder
But my words like silent raindrops fell. And echoed in the wells of silence.
