Dragon vs GM Nakamura Analog Handicap Match

Discussion of anything and everything relating to chess playing software and machines.

Moderators: Harvey Williamson, Dann Corbit, hgm

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Chessqueen
Posts: 1129
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: Dragon vs GM Nakamura Analog Handicap Match

Post by Chessqueen » Fri Nov 20, 2020 3:54 pm

JJJ wrote:
Fri Nov 20, 2020 10:17 am
I do think now that Dragon will crush Penguin at Knight handicap at 1+1 ! Incredible performance from Kromodo.
I am NOT too sure about that Humans under pressure blunder, but a Knight is a Knight and if penguin is at least 2450 he might beat Komodo MCTS as long as he keep a fortress up to the point where Komodo start playing faster and faster and later start trading pieces, where the extra Knight become too much of an advantage and half of the games should be with White pieces and the other half with Black Pieces for Penguin

Chessqueen
Posts: 1129
Joined: Wed Sep 05, 2018 12:16 am
Full name: Nancy M Pichardo

Re: Dragon vs GM Nakamura Analog Handicap Match

Post by Chessqueen » Fri Nov 20, 2020 4:01 pm

lkaufman wrote:
Wed Nov 18, 2020 12:26 am
Dragon vs. my son Ray was drawn 2 to 2 at knight odds, same tc as the Naka match. So if the difference in the odds offsets the strength difference between Naka and Ray, the Dragon vs Naka match should be close. It should be noted that standard Dragon lost 1.5 to 0.5 to Ray, while MCTS Dragon won by 1.5 to 0.5. Of course that's not a statistically significant difference, just a hint that it's okay to use MCTS for at least some games.

Have your son play two games with White pieces and two with Black pieces against MCTS Dragon. https://ratings.fide.com/profile/2015056

Marcus9
Posts: 50
Joined: Mon May 13, 2019 4:05 pm
Full name: Marco Giorgio

Re: Dragon vs GM Nakamura Analog Handicap Match

Post by Marcus9 » Fri Nov 20, 2020 6:31 pm

Chessqueen wrote:
Fri Nov 20, 2020 3:54 pm
JJJ wrote:
Fri Nov 20, 2020 10:17 am
I do think now that Dragon will crush Penguin at Knight handicap at 1+1 ! Incredible performance from Kromodo.
I am NOT too sure about that Humans under pressure blunder, but a Knight is a Knight and if penguin is at least 2450 he might beat Komodo MCTS as long as he keep a fortress up to the point where Komodo start playing faster and faster and later start trading pieces, where the extra Knight become too much of an advantage and half of the games should be with White pieces and the other half with Black Pieces for Penguin
Andrew Tang is pretty good at bullet, however I think dragon would be favorite, lc0 11248 wins GM Daniel Naroditsky at knight odd in blitz (3'+2") and draw in slow blitz (5'+5"), We all agree that Dragon is better than lc0.
I find it interesting to use knight odds without going over but changing TC against opponents of different strengths.

These are the match that I consider balanced Dragon with knight odd (b1 or g1):
GM Hikaru Nakamura: 1'+1"
GM Andrew Tang: 2'+1"
2700 elo blitz: 3'+2"
2600 elo blitz: 5'+5"
2500 elo rapid: 10'+5"
2400 elo rapid: 15'+10"

Maybe I would see Dragon slightly favored in most of these matches, however GM Larry Kaufman can correct me.

lkaufman
Posts: 4426
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: Dragon vs GM Nakamura Analog Handicap Match

Post by lkaufman » Fri Nov 20, 2020 6:53 pm

Chessqueen wrote:
Fri Nov 20, 2020 4:31 am
Chessqueen wrote:
Fri Nov 20, 2020 4:22 am
dkappe wrote:
Fri Nov 20, 2020 1:52 am
I think streaming on his twitch stream was Nakamura’s choice.

That is true and reason for losing so bad, If Carlsen ever accept this type of challenge he would at least draw 4 to 4 the match. Plus people here post that the MCTS is at least 80 Elo weaker, that could be when you are matching against others engines, but versus Human the style of MCTS is at least 75 Elo stronger than the Komodo Dragon NNUE or the first version that played versus Nakamura and much much more aggressive which makes humans blunder more.
I will ask Jorge Sammour to play versus Komodo MCTS under the same condition and time control, he should at least draw 1 game without talking, of course my new system is only 8 cores, but based on his Elo difference it should compensate.
Do you have Dragon? There is an enormous difference from prior Komodo, both in standard and MCTS mode.
Komodo rules!

lkaufman
Posts: 4426
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: Dragon vs GM Nakamura Analog Handicap Match

Post by lkaufman » Fri Nov 20, 2020 7:00 pm

Marcus9 wrote:
Fri Nov 20, 2020 6:31 pm
Chessqueen wrote:
Fri Nov 20, 2020 3:54 pm
JJJ wrote:
Fri Nov 20, 2020 10:17 am
I do think now that Dragon will crush Penguin at Knight handicap at 1+1 ! Incredible performance from Kromodo.
I am NOT too sure about that Humans under pressure blunder, but a Knight is a Knight and if penguin is at least 2450 he might beat Komodo MCTS as long as he keep a fortress up to the point where Komodo start playing faster and faster and later start trading pieces, where the extra Knight become too much of an advantage and half of the games should be with White pieces and the other half with Black Pieces for Penguin
Andrew Tang is pretty good at bullet, however I think dragon would be favorite, lc0 11248 wins GM Daniel Naroditsky at knight odd in blitz (3'+2") and draw in slow blitz (5'+5"), We all agree that Dragon is better than lc0.
I find it interesting to use knight odds without going over but changing TC against opponents of different strengths.

These are the match that I consider balanced Dragon with knight odd (b1 or g1):
GM Hikaru Nakamura: 1'+1"
GM Andrew Tang: 2'+1"
2700 elo blitz: 3'+2"
2600 elo blitz: 5'+5"
2500 elo rapid: 10'+5"
2400 elo rapid: 15'+10"

Maybe I would see Dragon slightly favored in most of these matches, however GM Larry Kaufman can correct me.
Assuming you are referring to FIDE blitz and rapid ratings, not chess.com or other server ratings, I think you are pretty accurate here. I was also thinking that 2' + 1" vs. Tang would be fair. Perhaps I'll try to contact him.
Komodo rules!

User avatar
Nordlandia
Posts: 2695
Joined: Fri Sep 25, 2015 7:38 pm
Location: Sortland, Norway

Re: Dragon vs GM Nakamura Analog Handicap Match

Post by Nordlandia » Fri Nov 20, 2020 7:19 pm

Dragon MCTS beat Komodo 14.1 in standard play ?

lkaufman
Posts: 4426
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: Dragon vs GM Nakamura Analog Handicap Match

Post by lkaufman » Fri Nov 20, 2020 8:09 pm

Nordlandia wrote:
Fri Nov 20, 2020 7:19 pm
Dragon MCTS beat Komodo 14.1 in standard play ?
It should, although I'll have to run a match to be precise about that.
Komodo rules!

User avatar
Nordlandia
Posts: 2695
Joined: Fri Sep 25, 2015 7:38 pm
Location: Sortland, Norway

Re: Dragon vs GM Nakamura Analog Handicap Match

Post by Nordlandia » Fri Nov 20, 2020 8:19 pm

lkaufman wrote:
Fri Nov 20, 2020 8:09 pm
Nordlandia wrote:
Fri Nov 20, 2020 7:19 pm
Dragon MCTS beat Komodo 14.1 in standard play ?
It should, although I'll have to run a match to be precise about that.
Excellent, if it can be done.

Uri Blass
Posts: 8836
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Dragon vs GM Nakamura Analog Handicap Match

Post by Uri Blass » Fri Nov 20, 2020 11:21 pm

lkaufman wrote:
Fri Nov 20, 2020 7:00 pm
Marcus9 wrote:
Fri Nov 20, 2020 6:31 pm
Chessqueen wrote:
Fri Nov 20, 2020 3:54 pm
JJJ wrote:
Fri Nov 20, 2020 10:17 am
I do think now that Dragon will crush Penguin at Knight handicap at 1+1 ! Incredible performance from Kromodo.
I am NOT too sure about that Humans under pressure blunder, but a Knight is a Knight and if penguin is at least 2450 he might beat Komodo MCTS as long as he keep a fortress up to the point where Komodo start playing faster and faster and later start trading pieces, where the extra Knight become too much of an advantage and half of the games should be with White pieces and the other half with Black Pieces for Penguin
Andrew Tang is pretty good at bullet, however I think dragon would be favorite, lc0 11248 wins GM Daniel Naroditsky at knight odd in blitz (3'+2") and draw in slow blitz (5'+5"), We all agree that Dragon is better than lc0.
I find it interesting to use knight odds without going over but changing TC against opponents of different strengths.

These are the match that I consider balanced Dragon with knight odd (b1 or g1):
GM Hikaru Nakamura: 1'+1"
GM Andrew Tang: 2'+1"
2700 elo blitz: 3'+2"
2600 elo blitz: 5'+5"
2500 elo rapid: 10'+5"
2400 elo rapid: 15'+10"

Maybe I would see Dragon slightly favored in most of these matches, however GM Larry Kaufman can correct me.
Assuming you are referring to FIDE blitz and rapid ratings, not chess.com or other server ratings, I think you are pretty accurate here. I was also thinking that 2' + 1" vs. Tang would be fair. Perhaps I'll try to contact him.
If we talk about fide blitz rating then tang's fide blitz rating is only 2584 and he had in the past blitz rating of 2603

https://ratings.fide.com/profile/2059630

I wonder how he can do well at 2'+1'' if 2600 elo blitz needs 5'+5''

Marcus9
Posts: 50
Joined: Mon May 13, 2019 4:05 pm
Full name: Marco Giorgio

Re: Dragon vs GM Nakamura Analog Handicap Match

Post by Marcus9 » Sat Nov 21, 2020 12:18 am

Uri Blass wrote:
Fri Nov 20, 2020 11:21 pm
lkaufman wrote:
Fri Nov 20, 2020 7:00 pm
Marcus9 wrote:
Fri Nov 20, 2020 6:31 pm
Chessqueen wrote:
Fri Nov 20, 2020 3:54 pm
JJJ wrote:
Fri Nov 20, 2020 10:17 am
I do think now that Dragon will crush Penguin at Knight handicap at 1+1 ! Incredible performance from Kromodo.
I am NOT too sure about that Humans under pressure blunder, but a Knight is a Knight and if penguin is at least 2450 he might beat Komodo MCTS as long as he keep a fortress up to the point where Komodo start playing faster and faster and later start trading pieces, where the extra Knight become too much of an advantage and half of the games should be with White pieces and the other half with Black Pieces for Penguin
Andrew Tang is pretty good at bullet, however I think dragon would be favorite, lc0 11248 wins GM Daniel Naroditsky at knight odd in blitz (3'+2") and draw in slow blitz (5'+5"), We all agree that Dragon is better than lc0.
I find it interesting to use knight odds without going over but changing TC against opponents of different strengths.

These are the match that I consider balanced Dragon with knight odd (b1 or g1):
GM Hikaru Nakamura: 1'+1"
GM Andrew Tang: 2'+1"
2700 elo blitz: 3'+2"
2600 elo blitz: 5'+5"
2500 elo rapid: 10'+5"
2400 elo rapid: 15'+10"

Maybe I would see Dragon slightly favored in most of these matches, however GM Larry Kaufman can correct me.
Assuming you are referring to FIDE blitz and rapid ratings, not chess.com or other server ratings, I think you are pretty accurate here. I was also thinking that 2' + 1" vs. Tang would be fair. Perhaps I'll try to contact him.
If we talk about fide blitz rating then tang's fide blitz rating is only 2584 and he had in the past blitz rating of 2603

https://ratings.fide.com/profile/2059630

I wonder how he can do well at 2'+1'' if 2600 elo blitz needs 5'+5''
Maybe I have a distorted conception of Andrew Tang's strength, I thought it was just short of the best players in the world in bullet, Maybe I was wrong.
In that case I would propose something really interesting:
Andrew Tang will have to win a match of 8 games at different time controls, with bigger prizes in the fast ones, he will start from those, if he loses he will extend the time available.

For example:
Andrew plays 8 games 2'+1", if win (4.5+ points) The prize will be 400$, If he draws or loses he will play 8 games 3'+2", again, if win It will earn 200$, Otherwise he will play the last match 5'+5" with 100$ In case of victory and 50$ If draws last match.

Obviously these are indicative prizes, the Komodo team must choose the prizes they deem most right.

I think this type of match would go a long way in understanding how strength changes at TC

Post Reply