Stockfish play very decent giving pawn odds.........

[Chessqueen]

I FOUND THIS INTERESTING GAME ON YOUTUBE where stockfish gave Nakamura the b7 and the h7 pawn at time control of 45:00 so Stockfish is not that bad playing odds, but at this time control Komodo can probably give Nakamura the f2 and f7 pawn which stockfish can not handle it




Here is Komodo 9 giving Nakamura the f2 pawn

Here is Komodo 9 giving Nakamura the f7 pawn

[Chessqueen]

I FOUND THIS INTERESTING GAME ON YOUTUBE where stockfish gave Nakamura the b7 and the h7 pawn at time control of 45:00 so Stockfish is not that bad playing odds, but at this time control Komodo can probably give Nakamura the f2 and f7 pawn which stockfish can not handle it




Here is Komodo 9 giving Nakamura the f2 pawn which is a much tougher position to defend

Here is Komodo 9 giving Nakamura the f7 pawn which is a much tougher position to defend

Mr. Larry, what is so clear to notice here is that Stockfish team gave two of the easier pawns such as b7 and h7 NOT f2 or f7 pawns to GM Nakamura , also Komodo 9 which is 5 versions earlier than Komodo 14 and at least 200 Elo weaker had a tougher position than Stockfish to defend and nowadays Komodo 14 can probably give Nakamura the same b7 and h7 pawn with longer time control of game in 90 minutes. Also in your Next Odds vs Human GM can you get a narrator as efficient and entertainer as Jerry who Narrated the Stockfish versus Nakamura game

PS: You can also get Kingscrusher who is an excellent narrator of most of the TCEC top engines games on youtube, he would keep everybody on their feet and aware of the best moves, a very entertainer chess narrator.

[lkaufman]

I FOUND THIS INTERESTING GAME ON YOUTUBE where stockfish gave Nakamura the b7 and the h7 pawn at time control of 45:00 so Stockfish is not that bad playing odds, but at this time control Komodo can probably give Nakamura the f2 and f7 pawn which stockfish can not handle it




Here is Komodo 9 giving Nakamura the f2 pawn which is a much tougher position to defend

Here is Komodo 9 giving Nakamura the f7 pawn which is a much tougher position to defend

Mr. Larry, what is so clear to notice here is that Stockfish team gave two of the easier pawns such as b7 and h7 NOT f2 or f7 pawns to GM Nakamura , also Komodo 9 which is 5 versions earlier than Komodo 14 and at least 200 Elo weaker had a tougher position than Stockfish to defend and nowadays Komodo 14 can probably give Nakamura the same b7 and h7 pawn with longer time control of game in 90 minutes. Also in your Next Odds vs Human GM can you get a narrator as efficient and entertainer as Jerry who Narrated the Stockfish versus Nakamura game

PS: You can also get Kingscrusher who is an excellent narrator of most of the TCEC top engines games on youtube, he would keep everybody on their feet and aware of the best moves, a very entertainer chess narrator.

Stockfish gave Nakamura b7 in one game and h7 in another, not two pawns in the same game, so this was a very small handicap, clearly both were less than f7 handicap. Stockfish is very strong when giving small handicaps like one pawn, maybe 2 pawns as White depending on the pawns, but when the odds grow larger Stockfish grows relatively weaker (compared to Komodo MCTS or Lc0 11248). At small handicaps there isn't much difference between playing to postpone losing and playing to maximize chances, but when normal play will almost certainly lose there is a huge difference.

[MonteCarlo]

I guess we'll see more about how it handles two pawns soon:

https://www.chess.com/news/view/stockfi ... an-foreest

[Chessqueen]

I guess we'll see more about how it handles two pawns soon:

https://www.chess.com/news/view/stockfi ... an-foreest

If games from 1 thru 5 were the positions from games 6 thru 10 : Stockfish without c- and f- pawns the match could finish in the first 2 games since the f pawn if very critical for any GM rated over 2650 to give it away as odds, even if the time control was only Blitz with 5 minutes total time control and Stockfish use the most powerful commercial IBM compatible computer the Pangea III,[/size] https://www.gigabitmagazine.com/ai/ibm- ... ercomputer

[MonteCarlo]

I'm not sure about it being such a sure thing at 5+0, with all the tactical mistakes made at that TC.

I don't think you're going to find a GM willing to test that out, though, so it'll probably stay speculation.

[Chessqueen]

I'm not sure about it being such a sure thing at 5+0, with all the tactical mistakes made at that TC.

I don't think you're going to find a GM willing to test that out, though, so it'll probably stay speculation.

If games from 1 thru 5 were the positions from games 6 thru 10 : Stockfish without c- and f- pawns the match could finish in the first 2 games since the f pawn is very critical and any GM rated over 2650 will take advantage of Stockfish that is not programmed to maximize and complicate the position, and SF only prolong the lost which you will notice very soon, even if the time control was only Blitz with 5 minutes total time control the GM rated above 2650 will give Stockfish a very hard time with the f2 or f7 removed plus any other pawn and of course that if with that computer specified to play in a few days

PS: Now the only program that is programmed to take advantage of that type of odds which is any pawn plus the f pawn versus GM is Komodo at the moment

[MonteCarlo]

Repeating the claim doesn't make me any more convinced

What will be tested in a few days are those odds at 30+10, which will us nothing about 5+0.

I'm well aware that SF does not play optimally against humans with large odds.

I'm also well aware how many tactical mistakes even very strong humans make in blitz without increment, which is why I remain unconvinced that the result is so certain in 5+0.

[lkaufman]

Repeating the claim doesn't make me any more convinced

What will be tested in a few days are those odds at 30+10, which will us nothing about 5+0.

I'm well aware that SF does not play optimally against humans with large odds.

I'm also well aware how many tactical mistakes even very strong humans make in blitz without increment, which is why I remain unconvinced that the result is so certain in 5+0.

Well, allegedly the sponsor is so sure that Stockfish won't lose a single game even at 30' + 10" that he is wagering $100,000 vs. $5,000 on the proposition!! (Although I haven't seen this confirmed in print yet, only the 20 to 1 odds). That sounds like an insane bet to me, and I warned against it, but perhaps he has enough other reasons for doing this, perhaps it's not really a "wager", rather a publicity stunt, I don't know. To me it's more like an even proposition. I think Stockfish should win more games than it loses if they were just playing for match victory, but only needing one win is another matter entirely. Stockfish isn't too bad at this level of handicap, but I think that aside from Komodo certain NNs would do better, particularly Fat Fritz, although even it falls apart at knight odds. I played test games myself against Stockfish 11 on my fast 8 core laptop at four of the handicaps at 10' + 5", and while I lost 3 I drew and came very close to winning at c2/f2. I reached a double rook endgame with two solid extra connected passers but messed it up with only the increment to play on to allow a draw. But I actually think that the handicaps involving g2 are even worse for the computer than c2/f2 objectively, but c2/f2 may be the worst for Stockfish because it won't open 1.e4, which all the NNs will choose.

[Vinvin]

I guess we'll see more about how it handles two pawns soon:

https://www.chess.com/news/view/stockfi ... an-foreest

You have to set a good value for the contempt.