Opponents for handicap match dragon-humans

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

Moderators: hgm, Dann Corbit, Harvey Williamson

lkaufman
Posts: 5942
Joined: Sun Jan 10, 2010 6:15 am
Location: Maryland USA

Re: Opponents for handicap match dragon-humans

Post by lkaufman »

Fritz 0 wrote: Sun May 08, 2022 11:30 am I'm impressed how much difference a time control makes. As I remember, GM Ben Finegold (and you are not!) struggled at 15+10. Yesterday much lower rated players scored equally well, or maybe even better, at 90+30.
Well, you say he "struggled", but he did win 4 to 2, better than the 8 to 5 score of the humans yesterday. But his FIDE rating was about 280 above their average rating, so the long time control for the humans (together with the simul handicap and no ponder for Dragon) did help them by a bit over a class (200 elo). I would guess my own chess play improves about a class going from Rapid to Classical.
Komodo rules!
Uri Blass
Posts: 10100
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Opponents for handicap match dragon-humans

Post by Uri Blass »

I think that it is not logical to have no ponder by dragon.
It basically means that there may be times during the games when dragon does nothing because it is the human to move in all the games
or only calculate one or 2 games out of 7 because it is the human turn in the other games.

Considering the fact that there is no equal time control so humans use most of the time it seems that it is better to have ponder on with the same hardware.
lkaufman
Posts: 5942
Joined: Sun Jan 10, 2010 6:15 am
Location: Maryland USA

Re: Opponents for handicap match dragon-humans

Post by lkaufman »

Uri Blass wrote: Sun May 08, 2022 8:04 pm I think that it is not logical to have no ponder by dragon.
It basically means that there may be times during the games when dragon does nothing because it is the human to move in all the games
or only calculate one or 2 games out of 7 because it is the human turn in the other games.

Considering the fact that there is no equal time control so humans use most of the time it seems that it is better to have ponder on with the same hardware.
It's not clear to me. With hyperthreading on, it may be better for the processors not to share threads with other processors thinking about moves that may not be played. You may be right, but it's probably a close call. In any case, I'm not sure that having more time to think is even useful (beyond Rapid) when giving knight odds. I didn't use MCTS for this event; if I did then it is more likely that Ponder would pay off.
Komodo rules!
Uri Blass
Posts: 10100
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Opponents for handicap match dragon-humans

Post by Uri Blass »

lkaufman wrote: Sun May 08, 2022 9:28 pm
Uri Blass wrote: Sun May 08, 2022 8:04 pm I think that it is not logical to have no ponder by dragon.
It basically means that there may be times during the games when dragon does nothing because it is the human to move in all the games
or only calculate one or 2 games out of 7 because it is the human turn in the other games.

Considering the fact that there is no equal time control so humans use most of the time it seems that it is better to have ponder on with the same hardware.
It's not clear to me. With hyperthreading on, it may be better for the processors not to share threads with other processors thinking about moves that may not be played. You may be right, but it's probably a close call. In any case, I'm not sure that having more time to think is even useful (beyond Rapid) when giving knight odds. I didn't use MCTS for this event; if I did then it is more likely that Ponder would pay off.
The main advantage of pondering is not playing better moves but replying faster to the opponent move so the opponent has less time to think on the computer time.
It is not a big advantage considering the fact that humans use most of the time.