planned weekly komodo vs human series on chess.com

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

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
lkaufman
Posts: 3763
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: planned weekly komodo vs human series on chess.com

Post by lkaufman » Wed Aug 21, 2019 3:07 am

Chessqueen wrote:
Tue Aug 20, 2019 11:24 pm
lkaufman wrote:
Tue Aug 20, 2019 4:02 pm
Nordlandia wrote:
Tue Aug 20, 2019 1:55 pm
Maybe Komodo team might give chess960 + material odds a try. The disharmony of the pieces offered in chess960 might offset some material deficit.

Most top engines play chess960 equally good as standard setup.
Didn't you read the previous post? Two of the games will be precisely what you ask, chess960 with knight odds. The selected positions will be ones that allow for normal castling, so they qualify both as chess960 and as normal chess with an altered start position. We already know from past matches with FM John Meyer that you are correct, it is much harder for the human to win with knight odds in 960 than in normal chess. Ray is a much stronger player, so we'll see if this is again the case.
Your son is rated 2062 in Chess960, you should give him a Knight odds plus a move https://www.chess.com/member/mimsychess
Some of his most recent games are from 2007 http://www.chessgames.com/player/raymond_s_kaufman.html

This game was well played versus such higher player Le Quang Liem back in 2014 what was the Vietnamese super star rating ?
From June 15 to June 20 Le participated in the 2014 World Rapid and Blitz Championship in Dubai, UAE as the defending champion in blitz. At blitz Liem finished fourth with +12=4-5, behind the new champion Magnus Carlsen, Ian Nepomniachtchi and Hikaru Nakamura. At rapid Liem finished 19th with +7=4-4.[16] ://www.chessgames.com/perl/chessgame?gid=1776899
Ray has a chess.com blitz rating near 2400, which is rather good; I imagine the chess960 rating was based on a few games and has no correlation with his current strength. Liem was already a super-GM 5 years ago, I don't recall just how super. Ray has beaten several GMs in tournaments and tied for the Canadian Open championship last year. He is a stronger chessplayer than I am now, especially at rapid/blitz. Note that the chess960 positions will feature normal castling with the kings and rooks on normal squares, so it's kind of a hybrid between normal chess and 960. Too bad only 18 normal positions (36 knight odds positions) qualify (including the standard one).
Komodo rules!

User avatar
Nordlandia
Posts: 2477
Joined: Fri Sep 25, 2015 7:38 pm
Location: Sortland, Norway

Re: planned weekly komodo vs human series on chess.com

Post by Nordlandia » Wed Aug 21, 2019 3:49 am

lkaufman wrote:
Tue Aug 20, 2019 4:02 pm
Nordlandia wrote:
Tue Aug 20, 2019 1:55 pm
Maybe Komodo team might give chess960 + material odds a try. The disharmony of the pieces offered in chess960 might offset some material deficit.

Most top engines play chess960 equally good as standard setup.
Didn't you read the previous post? Two of the games will be precisely what you ask, chess960 with knight odds. The selected positions will be ones that allow for normal castling, so they qualify both as chess960 and as normal chess with an altered start position. We already know from past matches with FM John Meyer that you are correct, it is much harder for the human to win with knight odds in 960 than in normal chess. Ray is a much stronger player, so we'll see if this is again the case.
I don't get why the the king's has to be on the original square for allowing normal castling. That completely contradicts the idea with Chess960.

Uri Blass
Posts: 8611
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: planned weekly komodo vs human series on chess.com

Post by Uri Blass » Wed Aug 21, 2019 5:04 am

Nordlandia wrote:
Wed Aug 21, 2019 3:49 am
lkaufman wrote:
Tue Aug 20, 2019 4:02 pm
Nordlandia wrote:
Tue Aug 20, 2019 1:55 pm
Maybe Komodo team might give chess960 + material odds a try. The disharmony of the pieces offered in chess960 might offset some material deficit.

Most top engines play chess960 equally good as standard setup.
Didn't you read the previous post? Two of the games will be precisely what you ask, chess960 with knight odds. The selected positions will be ones that allow for normal castling, so they qualify both as chess960 and as normal chess with an altered start position. We already know from past matches with FM John Meyer that you are correct, it is much harder for the human to win with knight odds in 960 than in normal chess. Ray is a much stronger player, so we'll see if this is again the case.
I don't get why the the king's has to be on the original square for allowing normal castling. That completely contradicts the idea with Chess960.
because if the king is not in the original square the interface does not understand castling.
It is not programmed for FRC.

i do not think that king in the original square completely contradicts the idea with chess960.

In part of the positions the king is in the original square.

lkaufman
Posts: 3763
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: planned weekly komodo vs human series on chess.com

Post by lkaufman » Wed Aug 21, 2019 6:16 am

Uri Blass wrote:
Wed Aug 21, 2019 5:04 am
Nordlandia wrote:
Wed Aug 21, 2019 3:49 am
lkaufman wrote:
Tue Aug 20, 2019 4:02 pm
Nordlandia wrote:
Tue Aug 20, 2019 1:55 pm
Maybe Komodo team might give chess960 + material odds a try. The disharmony of the pieces offered in chess960 might offset some material deficit.

Most top engines play chess960 equally good as standard setup.
Didn't you read the previous post? Two of the games will be precisely what you ask, chess960 with knight odds. The selected positions will be ones that allow for normal castling, so they qualify both as chess960 and as normal chess with an altered start position. We already know from past matches with FM John Meyer that you are correct, it is much harder for the human to win with knight odds in 960 than in normal chess. Ray is a much stronger player, so we'll see if this is again the case.
I don't get why the the king's has to be on the original square for allowing normal castling. That completely contradicts the idea with Chess960.
because if the king is not in the original square the interface does not understand castling.
It is not programmed for FRC.

i do not think that king in the original square completely contradicts the idea with chess960.

In part of the positions the king is in the original square.
To be more precise, chess.com does support chess 960, but does not support modifying the start position for 960, so we had to use positions that are legal in both 960 and normal chess, of which there are just 18 (17 + normal). Anyway I think it's more fair to the human this way, as at least he is not going to lose games by forgetting the castling rules or the option to castle. I played two test games myself against Playkomodo on chess.com tonite, duplicating the conditions for Ray's games Wednesday evening of the two chess 960 games, and I won the first rather routinely, simplying to a winning endgame (although I used almost all my time to do it). The second game I made a mistake on move 1 (!), but it wasn't critical and I reached the endgame still up knight for pawn. Komodo managed to sac a knight for my last pawn and the game was drawn with my two knights vs Komodo's sufficiently advanced one pawn. Ray is at least a hundred elo stronger than I am now in fast chess, so he should do well in the four knight odds games, including the two frc games.
I think frc would be much better if all positions allowed normal castling, but 18 positions are not nearly enough to prevent players from memorizing significant opening theory on each one. Maybe just requiring the king to be on e1/e8 would be a reasonable compromise.
Komodo rules!

User avatar
Nordlandia
Posts: 2477
Joined: Fri Sep 25, 2015 7:38 pm
Location: Sortland, Norway

Re: planned weekly komodo vs human series on chess.com

Post by Nordlandia » Wed Aug 21, 2019 6:24 am

Maybe Chess.com should update their setup mechanism.

The way i see it, if FRC is to be combined with material odds onwards. Making it as confusing as possible with randomization should in theory be better for the engine side which in return should allow for slightly higher odds allocated to the human side.

lkaufman
Posts: 3763
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: planned weekly komodo vs human series on chess.com

Post by lkaufman » Wed Aug 21, 2019 6:48 am

Nordlandia wrote:
Wed Aug 21, 2019 6:24 am
Maybe Chess.com should update their setup mechanism.

The way i see it, if FRC is to be combined with material odds onwards. Making it as confusing as possible with randomization should in theory be better for the engine side which in return should allow for slightly higher odds allocated to the human side.
That is true, but not clearly desirable. I'd rather it be closer to normal chess, just without the possibility of playing moves by rote without thinking.
My conclusion from my test games is that I need to set Contempt higher for Ray; some changes we made seem to require higher contempt for handicap games than before. So if Komodo does beat Ray in these 960 knight odds games, it will probably be due to raising Contempt.
Komodo rules!

lkaufman
Posts: 3763
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: planned weekly komodo vs human series on chess.com

Post by lkaufman » Wed Aug 21, 2019 7:36 pm

It's scheduled to start in a bit under 1 1/2 hours. I'm planning to raise Contempt to 200 for the knight odds games but only 100 for the other two handicaps, where material plays a lesser role in the handicap.
Komodo rules!

Vinvin
Posts: 4395
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: planned weekly komodo vs human series on chess.com

Post by Vinvin » Wed Aug 21, 2019 9:11 pm

lkaufman wrote:
Wed Aug 21, 2019 7:36 pm
It's scheduled to start in a bit under 1 1/2 hours. I'm planning to raise Contempt to 200 for the knight odds games but only 100 for the other two handicaps, where material plays a lesser role in the handicap.
How to find your game on chess.com ?

Henk
Posts: 5833
Joined: Mon May 27, 2013 8:31 am

Re: planned weekly komodo vs human series on chess.com

Post by Henk » Wed Aug 21, 2019 9:18 pm

F. Bluemers wrote:
Mon Aug 19, 2019 2:50 pm
Nay Lin Tun wrote:
Mon Aug 19, 2019 4:50 am
Father wrote:
Mon Aug 19, 2019 3:46 am
Thanks you very much Larry.

I appreciate the attention and deference you have with me; Being very close to reaching sixty years of age, we discovered the scope of that proverb that teaches: "Vanity of vanities and everything is vanity." I am very clear that Providence has wanted to place me in a wonderful place on stage. of chess: "Being human without titles will show the weaknesses of chess computers to satiety, will be despised by many and never recognized by the owners of the truth of the world." This sentence also contains vanity, due to that everything in this life is vanity and only vanity. For now Larry I will continue until God wants to play my game.
Best regards,
Pablo
I wonder, who has the bigger crazy Ego! ARB vs this dude!
Lyudmil Tsvetkov
Why have all interesting people left. It's a shame. No tolerance.

lkaufman
Posts: 3763
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: planned weekly komodo vs human series on chess.com

Post by lkaufman » Wed Aug 21, 2019 9:38 pm

Vinvin wrote:
Wed Aug 21, 2019 9:11 pm
lkaufman wrote:
Wed Aug 21, 2019 7:36 pm
It's scheduled to start in a bit under 1 1/2 hours. I'm planning to raise Contempt to 200 for the knight odds games but only 100 for the other two handicaps, where material plays a lesser role in the handicap.
How to find your game on chess.com ?
Observe "playkomodo" or "Mimsychess" (Ray). First game (b1 odds) drawn by repetition. Ray kept his extra knight, but with little time left besides the increment and no simple plan to progress he repeated.
Komodo rules!

Post Reply