Komodo vs. GM Lenderman match; "Almost a knight"

[lkaufman]

The time control was 10'+5" Jorge analyzed another way that the game could had gone but he was running out of time.

My main question is since GM Lenderman is about 100+ rating over Jorge Sammour is the difference in speed of your 32 cores will give Komodo 13 another 150 rating points than using my computer?

Well, yes Komodo will gain at least another 150 elo, probably more, but it's not that simple, because if both sides play 150 elo stronger then the odds-receiver should score more, since reducing the error rate for both players will obviously favor the one getting the handicap. So if Komodo and Sammour are really equal on your computer under the Lenderman conditions, then we might be in trouble. But we don't know that, as your first game was played with Contempt 16, and the second with bishop for pawn handicap. When you say he "analyzed" another continuation after the game, do you really mean he "played another game against Komodo from move 7". I mean if he analyzed it with Komodo on both sides surely Black would have won, and also the same if he analyzed it for both sides without Komodo.
If you can tell me the nodes per second Komodo gets in the initial position (normal chess, no handicap) on your computer with Threads set to whatever you used during this game, I can give you a much more accurate answer to how many elo stronger it will play on the Threadripper.

I am only getting 1545 kN/s in the initial position, No he analyzed it against Komodo 13 from move 7 on.........

On my 8 core laptop I get 13,000 kn/s in in initial position, on my 32 core Threadripper I get 57,500 KN/s (using 32 Threads; the count goes up to 81,000 kn/s if I use 64 but using more threads than cores artificially inflates the node count so I would use the lower figure). There is some wasting of nodes when running a lot of cores so let's say my threadripper is maybe close to 32 times as fast as your machine, so five doublings. The general rule for rapid games is something like 60 elo per doubling, so that would be 300 elo. Maybe 300 elo for Komodo offsets 150 elo for the human in a handicap game, hard to say. So perhaps Komodo on my threadripper vs Lenderman might do about as well as it does on your machine vs. Sammour. But we will have a newer, stronger Komodo, and the option to use MCTS or even an NN version if we think it will help. Still, based on your limited data and his comments, Komodo is probably an underdog in the overall Lenderman match, especially in the FRC games at full knight odds.
By the way, the analysis you quote is a bit strange, in that it is hard to imagine any strong player, human or engine, not playing 14...b6 but allowing 15.axb7+. Perhaps he wasn't trying to make best moves for Black for some reason. Or were you perhaps playing some or all of the Black moves yourself in this analysis?

I decided to play versus Komodo 13.3 by simply reversing the Knights and Bishops and removing the Knight from c1, since and I managed to forced Komodo 13.3 to NOT castle , I am showing you this few moves because it can throw away any opening theories where opening is useless but anyway my next move was going to be 10.d4 e4! which I believe it give me control of the center with pawns and I already prevented Komodo from castling. I will not continue this game because Komodo is too strong for me and at the end will beat me, but any IM or player over 2250 in my opinion using my computer that only calculate 1545 kN/s will give Komodo 13.3 a tough time. Are there any players in that rang like Uri who would like to continue this game? I know that at the end Komodo 13.3 will beat me in time or eventually even up the material but I manage to force it to NOT castle and I have control of the Center so far. Time control is Blitz:15'+10

Well I was so excited about my position that I called my friend Mauricio Ruiz rated 2257 told him to set up the position after what I was thinking of doing 10.....e4! and offered him $25.00 to finish it for me, he looked at it and said after 2 minutes of looking that he thinks that he can convert and Komodo 13.3 Resigned

[pgn]
[Event "Blitz:15'+10""]
[Site "MyTown"]
[Date "????.??.??"]
[Round "?"]
[White "Komodo 13.3 64-bit"]
[Black "Mauricio Ruiz"]
[Result "*"]
[SetUp "1"]
[FEN "rbnqknbr/pppppppp/8/8/8/8/PPPPPPPP/RB1QKNBR w KQkq - 0 1"]
[PlyCount "19"]
[TimeControl "900+10"]


1. c3 {[%emt 0:00:57]} e5 {[%emt 0:01:12]} 2. Ng3 {[%emt 0:00:53]} c6 {
[%emt 0:00:43]} 3. f3 {[%emt 0:00:24]} Ne6 {[%emt 0:00:22]} 4. Bf2 {
[%emt 0:00:20]} Bc7 {[%emt 0:00:48]} 5. a4 {[%emt 0:01:16]} h5 {[%emt 0:00:54]}
6. h4 {[%emt 0:02:18]} d5 {[%emt 0:00:08]} 7. b3 {[%emt 0:01:28]} Nd6 {
[%emt 0:01:18]} 8. Kf1 {[%emt 0:00:36]} g6 {[%emt 0:00:57]} 9. e3 {
[%emt 0:01:31]} f5 {[%emt 0:01:14]} 10. d4 {[%emt 0:00:23]} e4 {[%emt 0:00:00]}
11. f4 {[%emt 0:00:18]} Qf6 {[%emt 0:00:22]} 12. Bc2 {[%emt 0:00:34]} O-O-O {
[%emt 0:01:23]} 13. Kg1 {[%emt 0:00:12]} Qg7 {[%emt 0:01:43]} 14. Nf1 {
[%emt 0:00:50]} Qh6 {[%emt 0:00:06]} 15. Rh3 {[%emt 0:00:47]} Nxf4 {
[%emt 0:00:21]} 16. exf4 {[%emt 0:00:07]} Qxf4 {[%emt 0:00:06]} 17. Be3 {
[%emt 0:00:06]} Qg4 {[%emt 0:00:05]} 18. Qd2 {[%emt 0:00:06]} f4 {
[%emt 0:00:00]} 19. Bxf4 {[%emt 0:00:08]} Nc4 {[%emt 0:00:04]} 20. bxc4 {
[%emt 0:00:06]} Bxf4 {[%emt 0:00:44]} 21. Qe1 {[%emt 0:00:37]} Qd7 {
[%emt 0:00:17]} 22. Rb1 {[%emt 0:00:18]} Kb8 {[%emt 0:00:34]} 23. Ng3 {
[%emt 0:00:36]} Be6 {[%emt 0:00:20]} 24. Ne2 {[%emt 0:00:05]} g5 {
[%emt 0:00:38]} 25. Nxf4 {[%emt 0:00:00]} gxf4 {[%emt 0:00:25]} 26. Rh2 {
[%emt 0:00:10]} Rhg8 {[%emt 0:00:14]} 27. cxd5 {[%emt 0:00:17]} cxd5 {
[%emt 0:00:20]} 28. Rb2 {[%emt 0:00:40]} f3 {[%emt 0:02:35]} 29. Kh1 {
[%emt 0:00:00]} fxg2+ {[%emt 0:00:33]} 30. Rxg2 {[%emt 0:00:00]} Rxg2 {
[%emt 0:00:26]} 31. Kxg2 {[%emt 0:00:00]} Rg8+ {[%emt 0:00:18]} 32. Kh1 {
[%emt 0:00:00]} Bg4 {[%emt 0:00:35]} 33. Qg3+ {[%emt 0:00:00]} Ka8 {
[%emt 0:01:27]} 0-1[/pgn]

I strongly believe that even in Chess960 a Knight odds is too much and GM Lenderman has to give up one of his pawn b7, c7, d7, or e7 otherwise he will have it too easy, NO MATTER if larry Kaufman system is 32 times faster or even 64 times. Now thinking when Bobby Fischer said back in the 70's that he could had given any female player a Knight odds, Fischer was too Show off and we missed that opportunity for a rich person to make at least an offer from $20,000 thru $30,000 to see Fischer Ego being crashed I strongly believe that any FIDE 2225 thru 2275 can crash the champion Carlsen with a knight odds

I agree with most of your conclusions, however the game you show doesn't really show us too much, because there was at least one blunder by White due to the weak hardware, and because the position chosen was worse for White than necessary, remember in the 960 games I can choose among the 34 positions so I won't pick ones this bad for White. The 960 games should indeed be the worst for Komodo without the pawn or the castling rights as compensation for the knight, but playing with these conditions Komodo has performed around 2400 or so in test games so 1 out of 4 is a reasonable forecast for those games. Of course a 2250 FIDE would win a match from Carlsen at knight odds at 15' + 10", but one (my son) has already lost such a match to Komodo on the threadripper, that's a whole different level of difficulty. I'm pretty sure that Fischer was thinking of blitz when he made his knight-odds claim, or else assuming an American opponent (presumably Lisa Lane).

Show me two of the toughest Chess960 among those 34 positions, because Mauricio Ruiz would like to play two more games, and also what King safety settings should I use for Komodo 13.3 to protect his King better, since as he told me he is already up by a knight and his strategy is to get either his extra knight or Bishop close to Komodo KING and sacrifice it for two pawns and with that he break the king safety of Komodo 13.3 as shown in the previous game. Therefore what King safety setting is protect Komodo the best, and display two strong Chess960 positions for tomorrow games against Mauricio Ruiz FIDE 2257.
http://ratings.fide.com/card.phtml?event=2010593

Well, I don't want to reveal the positions for the match, but in general positions with the two Black knights starting on the same color squares are more difficult for Black than those with knights on opposite color squares. So pick any two of those. If you are not sure of how many cores you have or how many threads to use, just try setting Threads to 2, see if it runs properly, and then check kn/s in the normal initial position. If it goes above 2000, I think you can safely assume that you'll do better using 2 for Threads rather than one. Regarding King Safety, default is 84; a higher value would make it more willing for either side to sacrifice material to expose the king, and more reluctant for either side to allow this. So I think you want a higher value, at least raise it to 100, even a bit more if you are really worried about this, although too high a value might make Komodo go crazy with sacrifices itself. Maybe if you do all three of these things you'll have good chances.

Well I have these positions that I believe will be great against Mauricio Ruiz; but I want you to let me know if any of these positions is more difficult for a 2250+ FIDE rated to play versus Komodo 13.3. I will be pitting Komodo 13.3 versus Mauricio Ruiz tomorrow and I already figure tha with King safety set to 100 is better than 84 since I tested the previous game and Komodo avoided the immediate Attack to the King even after the Sacrifice of the Black Knight for the two pawns and followed by another Sacrifice that Komodo 13.3 simply felt for it.

1st because the bishop do not have a diagonal that can be used immediately
2nd both Black Knights are on the same color squares and are placed on the edge of the board
3rd the Rook are also not in the usual squares


Which two Chess960 positions could give Mauricio that toughest chess960 match and Why ?

=================================================================================================================
[d]qnrkbrbn/pppppppp/8/8/8/8/PPPPPPPP/Q1RKBRBN w - - 0 1

================================================================================================================
[d]qrknbrbn/pppppppp/8/8/8/8/PPPPPPPP/QRKNBRBN w - - 0 1
----------------------------------------------------------------------------------------------------------------------------------------------------------------
[d]nrqkbrnb/pppppppp/8/8/8/8/PPPPPPPP/1RQKBRNB w - - 0 1


Well I have these positions that I believe will be great against Mauricio Ruiz; but I want you to let me know if any of these positions is more difficult for a 2250+ FIDE rated to play versus Komodo 13.3. I will be pitting Komodo 13.3 versus Mauricio Ruiz tomorrow and I already figure tha with King safety set to 100 is better than 84 since I tested the previous game and Komodo avoided the immediate Attack to the King even after the Sacrifice of the Black Knight for the two pawns and followed by another Sacrifice that Komodo 13.3 simply felt for it.

1st because the bishop do not have a diagonal that can be used immediately
2nd both Black Knights are on the same color squares and are placed on the edge of the board
3rd the Rook are also not in the usual squares

Which two Chess960 positions could give Mauricio that toughest chess960 match and Why ?

Those are not even legal positions in chess960, never mind normal chess. In chess960 the two bishops for each player must always be on opposite color. For the Lenderman match I added the condition that the kings and rooks must also be on their normal squares, so neither he nor the viewers will have to think about the special castling rules of 960, they will be legal positions in normal chess. It's up to you whether you want to keep to the Lenderman conditions or not, but if you don't remember to select chess960 as the game type. Naturally your chances are better without the Lenderman rules, because he will be less familiar with how to play, whereas Komodo won't care, but be sure that you both know the chess 960 castling rules if you ignore the Lenderman condition. But please don't waste your time and money playing with bishops of the same color, that's neither chess nor chess960. I have no idea what chess960 position at knight odds (out of 1918 positions other than normal knight odds) is toughest for Black; of the 34 with the Lenderman rules I can't disclose my choices, but as long as you put the kings and rooks on normal squares, the bishops on opposite colors, and the (Black) knights on same color squares, it should be more difficult for him than normal knight odds as well as much less familiar to him. If you do it this way you don't have to stipulate chess 960 rules, the engine will play the same with either normal or 960 rules.

Well, you have Black to move in both positions; you obviously want it to be White to move in both. Also you forgot to set the castling rights for both sides, both should have normal castling, that's the whole point of using the normal king and rook positions so it will be legal to castle normally regardless of whether you use chess 960 rules or normal rules. If you correct those mistakes, then both positions are more difficult for Black than normal chess, it's just harder to get the Black pieces to good squares quickly than it is normally. But it's not a huge difference, it's still probably easier for Black than knight for pawn or knight for castling handicap. Remember to check whether two threads works well on your computer. That could be more important than these other points.

[Chessqueen]

The time control was 10'+5" Jorge analyzed another way that the game could had gone but he was running out of time.

My main question is since GM Lenderman is about 100+ rating over Jorge Sammour is the difference in speed of your 32 cores will give Komodo 13 another 150 rating points than using my computer?

Well, yes Komodo will gain at least another 150 elo, probably more, but it's not that simple, because if both sides play 150 elo stronger then the odds-receiver should score more, since reducing the error rate for both players will obviously favor the one getting the handicap. So if Komodo and Sammour are really equal on your computer under the Lenderman conditions, then we might be in trouble. But we don't know that, as your first game was played with Contempt 16, and the second with bishop for pawn handicap. When you say he "analyzed" another continuation after the game, do you really mean he "played another game against Komodo from move 7". I mean if he analyzed it with Komodo on both sides surely Black would have won, and also the same if he analyzed it for both sides without Komodo.
If you can tell me the nodes per second Komodo gets in the initial position (normal chess, no handicap) on your computer with Threads set to whatever you used during this game, I can give you a much more accurate answer to how many elo stronger it will play on the Threadripper.

I am only getting 1545 kN/s in the initial position, No he analyzed it against Komodo 13 from move 7 on.........

On my 8 core laptop I get 13,000 kn/s in in initial position, on my 32 core Threadripper I get 57,500 KN/s (using 32 Threads; the count goes up to 81,000 kn/s if I use 64 but using more threads than cores artificially inflates the node count so I would use the lower figure). There is some wasting of nodes when running a lot of cores so let's say my threadripper is maybe close to 32 times as fast as your machine, so five doublings. The general rule for rapid games is something like 60 elo per doubling, so that would be 300 elo. Maybe 300 elo for Komodo offsets 150 elo for the human in a handicap game, hard to say. So perhaps Komodo on my threadripper vs Lenderman might do about as well as it does on your machine vs. Sammour. But we will have a newer, stronger Komodo, and the option to use MCTS or even an NN version if we think it will help. Still, based on your limited data and his comments, Komodo is probably an underdog in the overall Lenderman match, especially in the FRC games at full knight odds.
By the way, the analysis you quote is a bit strange, in that it is hard to imagine any strong player, human or engine, not playing 14...b6 but allowing 15.axb7+. Perhaps he wasn't trying to make best moves for Black for some reason. Or were you perhaps playing some or all of the Black moves yourself in this analysis?

I decided to play versus Komodo 13.3 by simply reversing the Knights and Bishops and removing the Knight from c1, since and I managed to forced Komodo 13.3 to NOT castle , I am showing you this few moves because it can throw away any opening theories where opening is useless but anyway my next move was going to be 10.d4 e4! which I believe it give me control of the center with pawns and I already prevented Komodo from castling. I will not continue this game because Komodo is too strong for me and at the end will beat me, but any IM or player over 2250 in my opinion using my computer that only calculate 1545 kN/s will give Komodo 13.3 a tough time. Are there any players in that rang like Uri who would like to continue this game? I know that at the end Komodo 13.3 will beat me in time or eventually even up the material but I manage to force it to NOT castle and I have control of the Center so far. Time control is Blitz:15'+10

Well I was so excited about my position that I called my friend Mauricio Ruiz rated 2257 told him to set up the position after what I was thinking of doing 10.....e4! and offered him $25.00 to finish it for me, he looked at it and said after 2 minutes of looking that he thinks that he can convert and Komodo 13.3 Resigned

[pgn]
[Event "Blitz:15'+10""]
[Site "MyTown"]
[Date "????.??.??"]
[Round "?"]
[White "Komodo 13.3 64-bit"]
[Black "Mauricio Ruiz"]
[Result "*"]
[SetUp "1"]
[FEN "rbnqknbr/pppppppp/8/8/8/8/PPPPPPPP/RB1QKNBR w KQkq - 0 1"]
[PlyCount "19"]
[TimeControl "900+10"]


1. c3 {[%emt 0:00:57]} e5 {[%emt 0:01:12]} 2. Ng3 {[%emt 0:00:53]} c6 {
[%emt 0:00:43]} 3. f3 {[%emt 0:00:24]} Ne6 {[%emt 0:00:22]} 4. Bf2 {
[%emt 0:00:20]} Bc7 {[%emt 0:00:48]} 5. a4 {[%emt 0:01:16]} h5 {[%emt 0:00:54]}
6. h4 {[%emt 0:02:18]} d5 {[%emt 0:00:08]} 7. b3 {[%emt 0:01:28]} Nd6 {
[%emt 0:01:18]} 8. Kf1 {[%emt 0:00:36]} g6 {[%emt 0:00:57]} 9. e3 {
[%emt 0:01:31]} f5 {[%emt 0:01:14]} 10. d4 {[%emt 0:00:23]} e4 {[%emt 0:00:00]}
11. f4 {[%emt 0:00:18]} Qf6 {[%emt 0:00:22]} 12. Bc2 {[%emt 0:00:34]} O-O-O {
[%emt 0:01:23]} 13. Kg1 {[%emt 0:00:12]} Qg7 {[%emt 0:01:43]} 14. Nf1 {
[%emt 0:00:50]} Qh6 {[%emt 0:00:06]} 15. Rh3 {[%emt 0:00:47]} Nxf4 {
[%emt 0:00:21]} 16. exf4 {[%emt 0:00:07]} Qxf4 {[%emt 0:00:06]} 17. Be3 {
[%emt 0:00:06]} Qg4 {[%emt 0:00:05]} 18. Qd2 {[%emt 0:00:06]} f4 {
[%emt 0:00:00]} 19. Bxf4 {[%emt 0:00:08]} Nc4 {[%emt 0:00:04]} 20. bxc4 {
[%emt 0:00:06]} Bxf4 {[%emt 0:00:44]} 21. Qe1 {[%emt 0:00:37]} Qd7 {
[%emt 0:00:17]} 22. Rb1 {[%emt 0:00:18]} Kb8 {[%emt 0:00:34]} 23. Ng3 {
[%emt 0:00:36]} Be6 {[%emt 0:00:20]} 24. Ne2 {[%emt 0:00:05]} g5 {
[%emt 0:00:38]} 25. Nxf4 {[%emt 0:00:00]} gxf4 {[%emt 0:00:25]} 26. Rh2 {
[%emt 0:00:10]} Rhg8 {[%emt 0:00:14]} 27. cxd5 {[%emt 0:00:17]} cxd5 {
[%emt 0:00:20]} 28. Rb2 {[%emt 0:00:40]} f3 {[%emt 0:02:35]} 29. Kh1 {
[%emt 0:00:00]} fxg2+ {[%emt 0:00:33]} 30. Rxg2 {[%emt 0:00:00]} Rxg2 {
[%emt 0:00:26]} 31. Kxg2 {[%emt 0:00:00]} Rg8+ {[%emt 0:00:18]} 32. Kh1 {
[%emt 0:00:00]} Bg4 {[%emt 0:00:35]} 33. Qg3+ {[%emt 0:00:00]} Ka8 {
[%emt 0:01:27]} 0-1[/pgn]

I strongly believe that even in Chess960 a Knight odds is too much and GM Lenderman has to give up one of his pawn b7, c7, d7, or e7 otherwise he will have it too easy, NO MATTER if larry Kaufman system is 32 times faster or even 64 times. Now thinking when Bobby Fischer said back in the 70's that he could had given any female player a Knight odds, Fischer was too Show off and we missed that opportunity for a rich person to make at least an offer from $20,000 thru $30,000 to see Fischer Ego being crashed I strongly believe that any FIDE 2225 thru 2275 can crash the champion Carlsen with a knight odds

I agree with most of your conclusions, however the game you show doesn't really show us too much, because there was at least one blunder by White due to the weak hardware, and because the position chosen was worse for White than necessary, remember in the 960 games I can choose among the 34 positions so I won't pick ones this bad for White. The 960 games should indeed be the worst for Komodo without the pawn or the castling rights as compensation for the knight, but playing with these conditions Komodo has performed around 2400 or so in test games so 1 out of 4 is a reasonable forecast for those games. Of course a 2250 FIDE would win a match from Carlsen at knight odds at 15' + 10", but one (my son) has already lost such a match to Komodo on the threadripper, that's a whole different level of difficulty. I'm pretty sure that Fischer was thinking of blitz when he made his knight-odds claim, or else assuming an American opponent (presumably Lisa Lane).

Show me two of the toughest Chess960 among those 34 positions, because Mauricio Ruiz would like to play two more games, and also what King safety settings should I use for Komodo 13.3 to protect his King better, since as he told me he is already up by a knight and his strategy is to get either his extra knight or Bishop close to Komodo KING and sacrifice it for two pawns and with that he break the king safety of Komodo 13.3 as shown in the previous game. Therefore what King safety setting is protect Komodo the best, and display two strong Chess960 positions for tomorrow games against Mauricio Ruiz FIDE 2257.
http://ratings.fide.com/card.phtml?event=2010593

Well, I don't want to reveal the positions for the match, but in general positions with the two Black knights starting on the same color squares are more difficult for Black than those with knights on opposite color squares. So pick any two of those. If you are not sure of how many cores you have or how many threads to use, just try setting Threads to 2, see if it runs properly, and then check kn/s in the normal initial position. If it goes above 2000, I think you can safely assume that you'll do better using 2 for Threads rather than one. Regarding King Safety, default is 84; a higher value would make it more willing for either side to sacrifice material to expose the king, and more reluctant for either side to allow this. So I think you want a higher value, at least raise it to 100, even a bit more if you are really worried about this, although too high a value might make Komodo go crazy with sacrifices itself. Maybe if you do all three of these things you'll have good chances.

Well I have these positions that I believe will be great against Mauricio Ruiz; but I want you to let me know if any of these positions is more difficult for a 2250+ FIDE rated to play versus Komodo 13.3. I will be pitting Komodo 13.3 versus Mauricio Ruiz tomorrow and I already figure tha with King safety set to 100 is better than 84 since I tested the previous game and Komodo avoided the immediate Attack to the King even after the Sacrifice of the Black Knight for the two pawns and followed by another Sacrifice that Komodo 13.3 simply felt for it.

1st because the bishop do not have a diagonal that can be used immediately
2nd both Black Knights are on the same color squares and are placed on the edge of the board
3rd the Rook are also not in the usual squares


Which two Chess960 positions could give Mauricio that toughest chess960 match and Why ?

=================================================================================================================
[d]qnrkbrbn/pppppppp/8/8/8/8/PPPPPPPP/Q1RKBRBN w - - 0 1

================================================================================================================
[d]qrknbrbn/pppppppp/8/8/8/8/PPPPPPPP/QRKNBRBN w - - 0 1
----------------------------------------------------------------------------------------------------------------------------------------------------------------
[d]nrqkbrnb/pppppppp/8/8/8/8/PPPPPPPP/1RQKBRNB w - - 0 1


Well I have these positions that I believe will be great against Mauricio Ruiz; but I want you to let me know if any of these positions is more difficult for a 2250+ FIDE rated to play versus Komodo 13.3. I will be pitting Komodo 13.3 versus Mauricio Ruiz tomorrow and I already figure tha with King safety set to 100 is better than 84 since I tested the previous game and Komodo avoided the immediate Attack to the King even after the Sacrifice of the Black Knight for the two pawns and followed by another Sacrifice that Komodo 13.3 simply felt for it.

1st because the bishop do not have a diagonal that can be used immediately
2nd both Black Knights are on the same color squares and are placed on the edge of the board
3rd the Rook are also not in the usual squares

Which two Chess960 positions could give Mauricio that toughest chess960 match and Why ?

Those are not even legal positions in chess960, never mind normal chess. In chess960 the two bishops for each player must always be on opposite color. For the Lenderman match I added the condition that the kings and rooks must also be on their normal squares, so neither he nor the viewers will have to think about the special castling rules of 960, they will be legal positions in normal chess. It's up to you whether you want to keep to the Lenderman conditions or not, but if you don't remember to select chess960 as the game type. Naturally your chances are better without the Lenderman rules, because he will be less familiar with how to play, whereas Komodo won't care, but be sure that you both know the chess 960 castling rules if you ignore the Lenderman condition. But please don't waste your time and money playing with bishops of the same color, that's neither chess nor chess960. I have no idea what chess960 position at knight odds (out of 1918 positions other than normal knight odds) is toughest for Black; of the 34 with the Lenderman rules I can't disclose my choices, but as long as you put the kings and rooks on normal squares, the bishops on opposite colors, and the (Black) knights on same color squares, it should be more difficult for him than normal knight odds as well as much less familiar to him. If you do it this way you don't have to stipulate chess 960 rules, the engine will play the same with either normal or 960 rules.

Well, you have Black to move in both positions; you obviously want it to be White to move in both. Also you forgot to set the castling rights for both sides, both should have normal castling, that's the whole point of using the normal king and rook positions so it will be legal to castle normally regardless of whether you use chess 960 rules or normal rules. If you correct those mistakes, then both positions are more difficult for Black than normal chess, it's just harder to get the Black pieces to good squares quickly than it is normally. But it's not a huge difference, it's still probably easier for Black than knight for pawn or knight for castling handicap. Remember to check whether two threads works well on your computer. That could be more important than these other points.

Thanks for pointing that out that I had Black to move first, I selected a different position and we will be playing one more game, I believe that with the MCTS version Mauricio will have a very hard time, the contempt will be set to 150 King Safety to 100. I have confidence that Komodo will get him in Chess960, if NOT at least a draw will not be bad only $25.00

[Chessqueen]

The time control was 10'+5" Jorge analyzed another way that the game could had gone but he was running out of time.

My main question is since GM Lenderman is about 100+ rating over Jorge Sammour is the difference in speed of your 32 cores will give Komodo 13 another 150 rating points than using my computer?

Well, yes Komodo will gain at least another 150 elo, probably more, but it's not that simple, because if both sides play 150 elo stronger then the odds-receiver should score more, since reducing the error rate for both players will obviously favor the one getting the handicap. So if Komodo and Sammour are really equal on your computer under the Lenderman conditions, then we might be in trouble. But we don't know that, as your first game was played with Contempt 16, and the second with bishop for pawn handicap. When you say he "analyzed" another continuation after the game, do you really mean he "played another game against Komodo from move 7". I mean if he analyzed it with Komodo on both sides surely Black would have won, and also the same if he analyzed it for both sides without Komodo.
If you can tell me the nodes per second Komodo gets in the initial position (normal chess, no handicap) on your computer with Threads set to whatever you used during this game, I can give you a much more accurate answer to how many elo stronger it will play on the Threadripper.

I am only getting 1545 kN/s in the initial position, No he analyzed it against Komodo 13 from move 7 on.........

On my 8 core laptop I get 13,000 kn/s in in initial position, on my 32 core Threadripper I get 57,500 KN/s (using 32 Threads; the count goes up to 81,000 kn/s if I use 64 but using more threads than cores artificially inflates the node count so I would use the lower figure). There is some wasting of nodes when running a lot of cores so let's say my threadripper is maybe close to 32 times as fast as your machine, so five doublings. The general rule for rapid games is something like 60 elo per doubling, so that would be 300 elo. Maybe 300 elo for Komodo offsets 150 elo for the human in a handicap game, hard to say. So perhaps Komodo on my threadripper vs Lenderman might do about as well as it does on your machine vs. Sammour. But we will have a newer, stronger Komodo, and the option to use MCTS or even an NN version if we think it will help. Still, based on your limited data and his comments, Komodo is probably an underdog in the overall Lenderman match, especially in the FRC games at full knight odds.
By the way, the analysis you quote is a bit strange, in that it is hard to imagine any strong player, human or engine, not playing 14...b6 but allowing 15.axb7+. Perhaps he wasn't trying to make best moves for Black for some reason. Or were you perhaps playing some or all of the Black moves yourself in this analysis?

I decided to play versus Komodo 13.3 by simply reversing the Knights and Bishops and removing the Knight from c1, since and I managed to forced Komodo 13.3 to NOT castle , I am showing you this few moves because it can throw away any opening theories where opening is useless but anyway my next move was going to be 10.d4 e4! which I believe it give me control of the center with pawns and I already prevented Komodo from castling. I will not continue this game because Komodo is too strong for me and at the end will beat me, but any IM or player over 2250 in my opinion using my computer that only calculate 1545 kN/s will give Komodo 13.3 a tough time. Are there any players in that rang like Uri who would like to continue this game? I know that at the end Komodo 13.3 will beat me in time or eventually even up the material but I manage to force it to NOT castle and I have control of the Center so far. Time control is Blitz:15'+10

Well I was so excited about my position that I called my friend Mauricio Ruiz rated 2257 told him to set up the position after what I was thinking of doing 10.....e4! and offered him $25.00 to finish it for me, he looked at it and said after 2 minutes of looking that he thinks that he can convert and Komodo 13.3 Resigned

[pgn]
[Event "Blitz:15'+10""]
[Site "MyTown"]
[Date "????.??.??"]
[Round "?"]
[White "Komodo 13.3 64-bit"]
[Black "Mauricio Ruiz"]
[Result "*"]
[SetUp "1"]
[FEN "rbnqknbr/pppppppp/8/8/8/8/PPPPPPPP/RB1QKNBR w KQkq - 0 1"]
[PlyCount "19"]
[TimeControl "900+10"]


1. c3 {[%emt 0:00:57]} e5 {[%emt 0:01:12]} 2. Ng3 {[%emt 0:00:53]} c6 {
[%emt 0:00:43]} 3. f3 {[%emt 0:00:24]} Ne6 {[%emt 0:00:22]} 4. Bf2 {
[%emt 0:00:20]} Bc7 {[%emt 0:00:48]} 5. a4 {[%emt 0:01:16]} h5 {[%emt 0:00:54]}
6. h4 {[%emt 0:02:18]} d5 {[%emt 0:00:08]} 7. b3 {[%emt 0:01:28]} Nd6 {
[%emt 0:01:18]} 8. Kf1 {[%emt 0:00:36]} g6 {[%emt 0:00:57]} 9. e3 {
[%emt 0:01:31]} f5 {[%emt 0:01:14]} 10. d4 {[%emt 0:00:23]} e4 {[%emt 0:00:00]}
11. f4 {[%emt 0:00:18]} Qf6 {[%emt 0:00:22]} 12. Bc2 {[%emt 0:00:34]} O-O-O {
[%emt 0:01:23]} 13. Kg1 {[%emt 0:00:12]} Qg7 {[%emt 0:01:43]} 14. Nf1 {
[%emt 0:00:50]} Qh6 {[%emt 0:00:06]} 15. Rh3 {[%emt 0:00:47]} Nxf4 {
[%emt 0:00:21]} 16. exf4 {[%emt 0:00:07]} Qxf4 {[%emt 0:00:06]} 17. Be3 {
[%emt 0:00:06]} Qg4 {[%emt 0:00:05]} 18. Qd2 {[%emt 0:00:06]} f4 {
[%emt 0:00:00]} 19. Bxf4 {[%emt 0:00:08]} Nc4 {[%emt 0:00:04]} 20. bxc4 {
[%emt 0:00:06]} Bxf4 {[%emt 0:00:44]} 21. Qe1 {[%emt 0:00:37]} Qd7 {
[%emt 0:00:17]} 22. Rb1 {[%emt 0:00:18]} Kb8 {[%emt 0:00:34]} 23. Ng3 {
[%emt 0:00:36]} Be6 {[%emt 0:00:20]} 24. Ne2 {[%emt 0:00:05]} g5 {
[%emt 0:00:38]} 25. Nxf4 {[%emt 0:00:00]} gxf4 {[%emt 0:00:25]} 26. Rh2 {
[%emt 0:00:10]} Rhg8 {[%emt 0:00:14]} 27. cxd5 {[%emt 0:00:17]} cxd5 {
[%emt 0:00:20]} 28. Rb2 {[%emt 0:00:40]} f3 {[%emt 0:02:35]} 29. Kh1 {
[%emt 0:00:00]} fxg2+ {[%emt 0:00:33]} 30. Rxg2 {[%emt 0:00:00]} Rxg2 {
[%emt 0:00:26]} 31. Kxg2 {[%emt 0:00:00]} Rg8+ {[%emt 0:00:18]} 32. Kh1 {
[%emt 0:00:00]} Bg4 {[%emt 0:00:35]} 33. Qg3+ {[%emt 0:00:00]} Ka8 {
[%emt 0:01:27]} 0-1[/pgn]

I strongly believe that even in Chess960 a Knight odds is too much and GM Lenderman has to give up one of his pawn b7, c7, d7, or e7 otherwise he will have it too easy, NO MATTER if larry Kaufman system is 32 times faster or even 64 times. Now thinking when Bobby Fischer said back in the 70's that he could had given any female player a Knight odds, Fischer was too Show off and we missed that opportunity for a rich person to make at least an offer from $20,000 thru $30,000 to see Fischer Ego being crashed I strongly believe that any FIDE 2225 thru 2275 can crash the champion Carlsen with a knight odds

I agree with most of your conclusions, however the game you show doesn't really show us too much, because there was at least one blunder by White due to the weak hardware, and because the position chosen was worse for White than necessary, remember in the 960 games I can choose among the 34 positions so I won't pick ones this bad for White. The 960 games should indeed be the worst for Komodo without the pawn or the castling rights as compensation for the knight, but playing with these conditions Komodo has performed around 2400 or so in test games so 1 out of 4 is a reasonable forecast for those games. Of course a 2250 FIDE would win a match from Carlsen at knight odds at 15' + 10", but one (my son) has already lost such a match to Komodo on the threadripper, that's a whole different level of difficulty. I'm pretty sure that Fischer was thinking of blitz when he made his knight-odds claim, or else assuming an American opponent (presumably Lisa Lane).

Show me two of the toughest Chess960 among those 34 positions, because Mauricio Ruiz would like to play two more games, and also what King safety settings should I use for Komodo 13.3 to protect his King better, since as he told me he is already up by a knight and his strategy is to get either his extra knight or Bishop close to Komodo KING and sacrifice it for two pawns and with that he break the king safety of Komodo 13.3 as shown in the previous game. Therefore what King safety setting is protect Komodo the best, and display two strong Chess960 positions for tomorrow games against Mauricio Ruiz FIDE 2257.
http://ratings.fide.com/card.phtml?event=2010593

Well, I don't want to reveal the positions for the match, but in general positions with the two Black knights starting on the same color squares are more difficult for Black than those with knights on opposite color squares. So pick any two of those. If you are not sure of how many cores you have or how many threads to use, just try setting Threads to 2, see if it runs properly, and then check kn/s in the normal initial position. If it goes above 2000, I think you can safely assume that you'll do better using 2 for Threads rather than one. Regarding King Safety, default is 84; a higher value would make it more willing for either side to sacrifice material to expose the king, and more reluctant for either side to allow this. So I think you want a higher value, at least raise it to 100, even a bit more if you are really worried about this, although too high a value might make Komodo go crazy with sacrifices itself. Maybe if you do all three of these things you'll have good chances.

Well I have these positions that I believe will be great against Mauricio Ruiz; but I want you to let me know if any of these positions is more difficult for a 2250+ FIDE rated to play versus Komodo 13.3. I will be pitting Komodo 13.3 versus Mauricio Ruiz tomorrow and I already figure tha with King safety set to 100 is better than 84 since I tested the previous game and Komodo avoided the immediate Attack to the King even after the Sacrifice of the Black Knight for the two pawns and followed by another Sacrifice that Komodo 13.3 simply felt for it.

1st because the bishop do not have a diagonal that can be used immediately
2nd both Black Knights are on the same color squares and are placed on the edge of the board
3rd the Rook are also not in the usual squares


Which two Chess960 positions could give Mauricio that toughest chess960 match and Why ?

=================================================================================================================
[d]qnrkbrbn/pppppppp/8/8/8/8/PPPPPPPP/Q1RKBRBN w - - 0 1

================================================================================================================
[d]qrknbrbn/pppppppp/8/8/8/8/PPPPPPPP/QRKNBRBN w - - 0 1
----------------------------------------------------------------------------------------------------------------------------------------------------------------
[d]nrqkbrnb/pppppppp/8/8/8/8/PPPPPPPP/1RQKBRNB w - - 0 1


Well I have these positions that I believe will be great against Mauricio Ruiz; but I want you to let me know if any of these positions is more difficult for a 2250+ FIDE rated to play versus Komodo 13.3. I will be pitting Komodo 13.3 versus Mauricio Ruiz tomorrow and I already figure tha with King safety set to 100 is better than 84 since I tested the previous game and Komodo avoided the immediate Attack to the King even after the Sacrifice of the Black Knight for the two pawns and followed by another Sacrifice that Komodo 13.3 simply felt for it.

1st because the bishop do not have a diagonal that can be used immediately
2nd both Black Knights are on the same color squares and are placed on the edge of the board
3rd the Rook are also not in the usual squares

Which two Chess960 positions could give Mauricio that toughest chess960 match and Why ?

Those are not even legal positions in chess960, never mind normal chess. In chess960 the two bishops for each player must always be on opposite color. For the Lenderman match I added the condition that the kings and rooks must also be on their normal squares, so neither he nor the viewers will have to think about the special castling rules of 960, they will be legal positions in normal chess. It's up to you whether you want to keep to the Lenderman conditions or not, but if you don't remember to select chess960 as the game type. Naturally your chances are better without the Lenderman rules, because he will be less familiar with how to play, whereas Komodo won't care, but be sure that you both know the chess 960 castling rules if you ignore the Lenderman condition. But please don't waste your time and money playing with bishops of the same color, that's neither chess nor chess960. I have no idea what chess960 position at knight odds (out of 1918 positions other than normal knight odds) is toughest for Black; of the 34 with the Lenderman rules I can't disclose my choices, but as long as you put the kings and rooks on normal squares, the bishops on opposite colors, and the (Black) knights on same color squares, it should be more difficult for him than normal knight odds as well as much less familiar to him. If you do it this way you don't have to stipulate chess 960 rules, the engine will play the same with either normal or 960 rules.

Well, you have Black to move in both positions; you obviously want it to be White to move in both. Also you forgot to set the castling rights for both sides, both should have normal castling, that's the whole point of using the normal king and rook positions so it will be legal to castle normally regardless of whether you use chess 960 rules or normal rules. If you correct those mistakes, then both positions are more difficult for Black than normal chess, it's just harder to get the Black pieces to good squares quickly than it is normally. But it's not a huge difference, it's still probably easier for Black than knight for pawn or knight for castling handicap. Remember to check whether two threads works well on your computer. That could be more important than these other points.

Thanks for pointing that out that I had Black to move first, I selected a different position and we will be playing one more game, I believe that with the MCTS version Mauricio will have a very hard time, the contempt will be set to 150 King Safety to 100. I have confidence that Komodo will get him in Chess960, if NOT at least a draw will not be bad only $25.00

Here is the DRAW, but very exciting game

[pgn]
[Event "?"]
[Site "?"]
[Date "2020.04.28"]
[Round "?"]
[White "Komodo 13.3 64-bit"]
[Black "Mauricio Ruiz"]
[Result "1/2-1/2"]
[Annotator "ChessQueen"]
[SetUp "1"]
[FEN "rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1"]
[PlyCount "121"]

1. f4 {[%emt 0:00:53]} d5 {[%emt 0:01:19]} 2. d3 {[%emt 0:00:37]} c6 {
[%emt 0:00:45]} 3. e4 {[%emt 0:00:38]} dxe4 {[%emt 0:00:26]} 4. dxe4 {
[%emt 0:00:53]} f5 {[%emt 0:00:21]} 5. Qf2 {[%emt 0:00:44]} Nbd7 {
[%emt 0:00:42]} 6. exf5 {[%emt 0:01:04]} Nf6 {[%emt 0:00:19]} 7. O-O {
[%emt 0:00:32]} Bxf5 {[%emt 0:01:34]} 8. c3 {[%emt 0:00:00]} Bxb1 {
[%emt 0:00:39]} 9. Rxb1 {[%emt 0:00:32]} h5 {[%emt 0:00:03]} 10. Be3 {
[%emt 0:00:47]} N8d7 {[%emt 0:00:25]} 11. Bb3 {[%emt 0:00:31]} Nd5 {
[%emt 0:00:29]} 12. Rbd1 {[%emt 0:00:53]} e6 {[%emt 0:00:28]} 13. Bd4 {
[%emt 0:00:48]} Bf6 {[%emt 0:02:16]} 14. Bxa7 {[%emt 0:00:43]} b6 {
[%emt 0:00:30]} 15. Bxb6 {[%emt 0:00:01]} N7xb6 {[%emt 0:00:09]} 16. c4 {
[%emt 0:00:30]} Nxc4 {[%emt 0:00:13]} 17. Bxc4 {[%emt 0:00:25]} Qf7 {
[%emt 0:00:32]} 18. Kh1 {[%emt 0:00:01]} g6 {[%emt 0:00:18]} 19. Qc5 {
[%emt 0:00:50]} Qb7 {[%emt 0:00:46]} 20. Qf2 {[%emt 0:00:09]} Qxb2 {
[%emt 0:00:29]} 21. Qe1 {[%emt 0:00:58]} Kf7 {[%emt 0:00:19]} 22. Rb1 {
[%emt 0:00:11]} Qc3 {[%emt 0:00:52]} 23. Rb7+ {[%emt 0:00:25]} Be7 {
[%emt 0:00:37]} 24. Qxc3 {[%emt 0:00:02]} Nxc3 {[%emt 0:00:39]} 25. f5 {
[%emt 0:00:01]} gxf5 {[%emt 0:00:13]} 26. Rxf5+ {[%emt 0:00:14]} Ke8 {
[%emt 0:00:24]} 27. Bxe6 {[%emt 0:00:27]} Rh6 {[%emt 0:00:09]} 28. Re5 {
[%emt 0:00:30]} h4 {[%emt 0:00:00]} 29. a4 {[%emt 0:00:56]} Rxa4 {
[%emt 0:00:09]} 30. Rb8+ {[%emt 0:00:13]} Bd8 {[%emt 0:00:31]} 31. Bb3+ {
[%emt 0:00:00]} Kd7 {[%emt 0:00:09]} 32. Bxa4 {[%emt 0:00:17]} Nxa4 {
[%emt 0:00:17]} 33. Rb3 {[%emt 0:00:22]} Nb6 {[%emt 0:01:28]} 34. Kg1 {
[%emt 0:00:00]} Nd5 {[%emt 0:00:27]} 35. g3 {[%emt 0:00:37]} Bc7 {
[%emt 0:00:14]} 36. Rg5 {[%emt 0:00:16]} Bb6+ {[%emt 0:00:16]} 37. Kg2 {
[%emt 0:00:24]} h3+ {[%emt 0:00:14]} 38. Kf3 {[%emt 0:00:13]} Bg1 {
[%emt 0:00:12]} 39. Rb2 {[%emt 0:00:16]} Kd6 {[%emt 0:00:07]} 40. Ra2 {
[%emt 0:00:21]} Re6 {[%emt 0:00:19]} 41. Kg4 {[%emt 0:00:13]} Rh6 {
[%emt 0:00:05]} 42. Kf3 {[%emt 0:00:28]} Bd4 {[%emt 0:00:05]} 43. Ra4 {
[%emt 0:00:24]} Rf6+ {[%emt 0:00:14]} 44. Kg4 {[%emt 0:00:07]} Ne3+ {
[%emt 0:00:11]} 45. Kh5 {[%emt 0:00:10]} Bb2 {[%emt 0:00:06]} 46. Re4 {
[%emt 0:00:26]} Nf1 {[%emt 0:00:07]} 47. Kg4 {[%emt 0:00:27]} Nxh2+ {
[%emt 0:00:13]} 48. Kxh3 {[%emt 0:00:18]} Nf3 {[%emt 0:00:20]} 49. Rg8 {
[%emt 0:00:41]} c5 {[%emt 0:00:35]} 50. Ra4 {[%emt 0:00:25]} Nd4 {
[%emt 0:00:12]} 51. Rd8+ {[%emt 0:00:33]} Kc7 {[%emt 0:00:18]} 52. Rd5 {
[%emt 0:00:07]} Kc6 {[%emt 0:00:04]} 53. Rh5 {[%emt 0:00:24]} Kb5 {
[%emt 0:00:05]} 54. Ra2 {[%emt 0:00:15]} Bc3 {[%emt 0:00:13]} 55. Ra7 {
[%emt 0:00:14]} Nc6 {[%emt 0:00:19]} 56. Rb7+ {[%emt 0:00:20]} Kc4 {
[%emt 0:00:06]} 57. Rh4+ {[%emt 0:00:20]} Bd4 {[%emt 0:00:13]} 58. Kg2 {
[%emt 0:00:05]} Rf2+ {[%emt 0:00:08]} 59. Kh3 {[%emt 0:00:17]} Re2 {
[%emt 0:00:05]} 60. Rf4 {[%emt 0:00:22]} Ne5 {[%emt 0:00:23]} 61. Ra7 {
[%emt 0:00:19]} 1/2-1/2[/pgn]

[Chessqueen]

Looks like you picked a reasonable position and did everything properly. The final position was probably theoretically still losing for Komodo but probably too difficult for the human to win with just increment. Most of the moves looked either best or close to best on my 8 core laptop; only suspicious move was 8.c3?! but his answer was bad so it worked out ok. Did you determine whether you could use two threads or just the one? We still don't know whether MCTS actually helps in these knight (or knight for pawn) odds games vs. humans; overall results have been better, but not by enough to claim statistical significance. At least this game does suggest that the 960 factor makes the knight odds a little less easy for the human than normal knight odds.

Well I tried two threads and it was not any better but when I try 4 threads I saw an improvements up to 2200 kN/s, but I decided to use MCTS, the final position was losing for Komodo, but according to Mauricio he asked for a draw and I accepted, because at the very beginning he told me if I come close to only 20 seconds even if I am winning I will offer a draw because from experience the engine will play extremely fast without blundering but there is no way for me to beat the engine and finding the best moves with less than 20 seconds. Finally thought is that GM Lenderman will still need to remove on of his b7, c7,d7, or e7 pawn with Chess960 to make it competitive even with your super fast 32 cores system

Larry can you or your son play to us at least one game Versus your 32 core with this Chess960 Position. We would like to see how convincing your system WIN with this position using all that calculating power with the MCTS version or even a bigger surprise with your NN version, we only hear rumors but never seeing any game of your NN komodo yet. Or even better using this position pit your regular version against the MCTS or the NN version for comparison reason playing both side but without taking the knight off if it is engine vs engine , unless you use a weaker computer with black then keep this position against your 32 core monster computer just to compare:roll:

[d]rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1

[lkaufman]

Looks like you picked a reasonable position and did everything properly. The final position was probably theoretically still losing for Komodo but probably too difficult for the human to win with just increment. Most of the moves looked either best or close to best on my 8 core laptop; only suspicious move was 8.c3?! but his answer was bad so it worked out ok. Did you determine whether you could use two threads or just the one? We still don't know whether MCTS actually helps in these knight (or knight for pawn) odds games vs. humans; overall results have been better, but not by enough to claim statistical significance. At least this game does suggest that the 960 factor makes the knight odds a little less easy for the human than normal knight odds.

Well I tried two threads and it was not any better but when I try 4 threads I saw an improvements up to 2200 kN/s, but I decided to use MCTS, the final position was losing for Komodo, but according to Mauricio he asked for a draw and I accepted, because at the very beginning he told me if I come close to only 20 seconds even if I am winning I will offer a draw because from experience the engine will play extremely fast without blundering but there is no way for me to beat the engine and finding the best moves with less than 20 seconds. Finally thought is that GM Lenderman will still need to remove on of his b7, c7,d7, or e7 pawn with Chess960 to make it competitive even with your super fast 32 cores system

Larry can you or your son play to us at least one game Versus your 32 core with this Chess960 Position. We would like to see how convincing your system WIN with this position using all that calculating power with the MCTS version or even a bigger surprise with your NN version, we only hear rumors but never seeing any game of your NN komodo yet. Or even better using this position pit your regular version against the MCTS or the NN version for comparison reason playing both side but without taking the knight off if it is engine vs engine , unless you use a weaker computer with black then keep this position against your 32 core monster computer just to compare:roll:

[d]rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1

Yes, I can play a game against it on my 32 core myself from that position at 15' + 10" with the MCTS version. I lose more than I win at FRC knight odds with K on this hardware at this time limit, but the most frequent result is draw. We did play our last match vs. FM Larry Gilden with the NN version, so there are games with it posted on talkchess, but it's not in proper shape to use for a test right now. Playing normal K vs. MCTS from an even position is pointless, we already have thousands of games that show that regular K is about 80 elo stronger, it doesn't much matter whether it's normal chess or 960. Playing the 32 core against a 1 core at knight odds also won't work, as we have lots of tests that show that this sort of hardware advantage only offsets something like pawn and move odds, not nearly knight odds. Normal Komodo at bullet vs. Komodo Level 23 is reasonably close at knight odds, so you could also try having your friend play normal chess (15' + 10" with Komodo level 23 (not MCTS) to see who is stronger. Komodo level 23 will move instantly, so game takes only half the time of a knight odds game with full strength komodo.

[Chessqueen]

Looks like you picked a reasonable position and did everything properly. The final position was probably theoretically still losing for Komodo but probably too difficult for the human to win with just increment. Most of the moves looked either best or close to best on my 8 core laptop; only suspicious move was 8.c3?! but his answer was bad so it worked out ok. Did you determine whether you could use two threads or just the one? We still don't know whether MCTS actually helps in these knight (or knight for pawn) odds games vs. humans; overall results have been better, but not by enough to claim statistical significance. At least this game does suggest that the 960 factor makes the knight odds a little less easy for the human than normal knight odds.

Well I tried two threads and it was not any better but when I try 4 threads I saw an improvements up to 2200 kN/s, but I decided to use MCTS, the final position was losing for Komodo, but according to Mauricio he asked for a draw and I accepted, because at the very beginning he told me if I come close to only 20 seconds even if I am winning I will offer a draw because from experience the engine will play extremely fast without blundering but there is no way for me to beat the engine and finding the best moves with less than 20 seconds. Finally thought is that GM Lenderman will still need to remove on of his b7, c7,d7, or e7 pawn with Chess960 to make it competitive even with your super fast 32 cores system

Larry can you or your son play to us at least one game Versus your 32 core with this Chess960 Position. We would like to see how convincing your system WIN with this position using all that calculating power with the MCTS version or even a bigger surprise with your NN version, we only hear rumors but never seeing any game of your NN komodo yet. Or even better using this position pit your regular version against the MCTS or the NN version for comparison reason playing both side but without taking the knight off if it is engine vs engine , unless you use a weaker computer with black then keep this position against your 32 core monster computer just to compare:roll:

[d]rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1

Yes, I can play a game against it on my 32 core myself from that position at 15' + 10" with the MCTS version. I lose more than I win at FRC knight odds with K on this hardware at this time limit, but the most frequent result is draw. We did play our last match vs. FM Larry Gilden with the NN version, so there are games with it posted on talkchess, but it's not in proper shape to use for a test right now. Playing normal K vs. MCTS from an even position is pointless, we already have thousands of games that show that regular K is about 80 elo stronger, it doesn't much matter whether it's normal chess or 960. Playing the 32 core against a 1 core at knight odds also won't work, as we have lots of tests that show that this sort of hardware advantage only offsets something like pawn and move odds, not nearly knight odds. Normal Komodo at bullet vs. Komodo Level 23 is reasonably close at knight odds, so you could also try having your friend play normal chess (15' + 10" with Komodo level 23 (not MCTS) to see who is stronger. Komodo level 23 will move instantly, so game takes only half the time of a knight odds game with full strength komodo.

This will be interesting Can you play one game Vs the MCTS and another with the Regular komodo just to compare which will give you the hardest time ?

[Chessqueen]

Looks like you picked a reasonable position and did everything properly. The final position was probably theoretically still losing for Komodo but probably too difficult for the human to win with just increment. Most of the moves looked either best or close to best on my 8 core laptop; only suspicious move was 8.c3?! but his answer was bad so it worked out ok. Did you determine whether you could use two threads or just the one? We still don't know whether MCTS actually helps in these knight (or knight for pawn) odds games vs. humans; overall results have been better, but not by enough to claim statistical significance. At least this game does suggest that the 960 factor makes the knight odds a little less easy for the human than normal knight odds.

Well I tried two threads and it was not any better but when I try 4 threads I saw an improvements up to 2200 kN/s, but I decided to use MCTS, the final position was losing for Komodo, but according to Mauricio he asked for a draw and I accepted, because at the very beginning he told me if I come close to only 20 seconds even if I am winning I will offer a draw because from experience the engine will play extremely fast without blundering but there is no way for me to beat the engine and finding the best moves with less than 20 seconds. Finally thought is that GM Lenderman will still need to remove on of his b7, c7,d7, or e7 pawn with Chess960 to make it competitive even with your super fast 32 cores system

Larry can you or your son play to us at least one game Versus your 32 core with this Chess960 Position. We would like to see how convincing your system WIN with this position using all that calculating power with the MCTS version or even a bigger surprise with your NN version, we only hear rumors but never seeing any game of your NN komodo yet. Or even better using this position pit your regular version against the MCTS or the NN version for comparison reason playing both side but without taking the knight off if it is engine vs engine , unless you use a weaker computer with black then keep this position against your 32 core monster computer just to compare:roll:

[d]rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1

Yes, I can play a game against it on my 32 core myself from that position at 15' + 10" with the MCTS version. I lose more than I win at FRC knight odds with K on this hardware at this time limit, but the most frequent result is draw. We did play our last match vs. FM Larry Gilden with the NN version, so there are games with it posted on talkchess, but it's not in proper shape to use for a test right now. Playing normal K vs. MCTS from an even position is pointless, we already have thousands of games that show that regular K is about 80 elo stronger, it doesn't much matter whether it's normal chess or 960. Playing the 32 core against a 1 core at knight odds also won't work, as we have lots of tests that show that this sort of hardware advantage only offsets something like pawn and move odds, not nearly knight odds. Normal Komodo at bullet vs. Komodo Level 23 is reasonably close at knight odds, so you could also try having your friend play normal chess (15' + 10" with Komodo level 23 (not MCTS) to see who is stronger. Komodo level 23 will move instantly, so game takes only half the time of a knight odds game with full strength komodo.

This will be interesting Can you play one game Vs the MCTS and another with the Regular komodo just to compare which will give you the hardest time ?

The way I found out how many core my system has is by holding the Ctrl + Shift + Esc and I clicked on the Performance it told me 4 core. Now if your 32 core is giving you a hard time select one with either 16 or 8 Core, since my 4 core almost got Mauricio Ruiz with this position.

[lkaufman]

Looks like you picked a reasonable position and did everything properly. The final position was probably theoretically still losing for Komodo but probably too difficult for the human to win with just increment. Most of the moves looked either best or close to best on my 8 core laptop; only suspicious move was 8.c3?! but his answer was bad so it worked out ok. Did you determine whether you could use two threads or just the one? We still don't know whether MCTS actually helps in these knight (or knight for pawn) odds games vs. humans; overall results have been better, but not by enough to claim statistical significance. At least this game does suggest that the 960 factor makes the knight odds a little less easy for the human than normal knight odds.

Well I tried two threads and it was not any better but when I try 4 threads I saw an improvements up to 2200 kN/s, but I decided to use MCTS, the final position was losing for Komodo, but according to Mauricio he asked for a draw and I accepted, because at the very beginning he told me if I come close to only 20 seconds even if I am winning I will offer a draw because from experience the engine will play extremely fast without blundering but there is no way for me to beat the engine and finding the best moves with less than 20 seconds. Finally thought is that GM Lenderman will still need to remove on of his b7, c7,d7, or e7 pawn with Chess960 to make it competitive even with your super fast 32 cores system

Larry can you or your son play to us at least one game Versus your 32 core with this Chess960 Position. We would like to see how convincing your system WIN with this position using all that calculating power with the MCTS version or even a bigger surprise with your NN version, we only hear rumors but never seeing any game of your NN komodo yet. Or even better using this position pit your regular version against the MCTS or the NN version for comparison reason playing both side but without taking the knight off if it is engine vs engine , unless you use a weaker computer with black then keep this position against your 32 core monster computer just to compare:roll:

[d]rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1

Yes, I can play a game against it on my 32 core myself from that position at 15' + 10" with the MCTS version. I lose more than I win at FRC knight odds with K on this hardware at this time limit, but the most frequent result is draw. We did play our last match vs. FM Larry Gilden with the NN version, so there are games with it posted on talkchess, but it's not in proper shape to use for a test right now. Playing normal K vs. MCTS from an even position is pointless, we already have thousands of games that show that regular K is about 80 elo stronger, it doesn't much matter whether it's normal chess or 960. Playing the 32 core against a 1 core at knight odds also won't work, as we have lots of tests that show that this sort of hardware advantage only offsets something like pawn and move odds, not nearly knight odds. Normal Komodo at bullet vs. Komodo Level 23 is reasonably close at knight odds, so you could also try having your friend play normal chess (15' + 10" with Komodo level 23 (not MCTS) to see who is stronger. Komodo level 23 will move instantly, so game takes only half the time of a knight odds game with full strength komodo.

This will be interesting Can you play one game Vs the MCTS and another with the Regular komodo just to compare which will give you the hardest time ?

The way I found out how many core my system has is by holding the Ctrl + Shift + Esc and I clicked on the Performance it told me 4 core. Now if your 32 core is giving you a hard time select one with either 16 or 8 Core, since my 4 core almost got Mauricio Ruiz with this position.

I only had time for the MCTS game tonite. I could of course have easily made a draw by playing ...g4 at almost any point to lock up the game, but drawing Komodo at knight odds is no big achievement, the challenge is to win, so I went for it and lost. Maybe I can play another game without MCTS tomorrow, but it may tell us more about what I have learned from the game than whether MCTS helps or not.

[Event "Live Chess - FRCNODDS"]
[Site "Chess.com"]
[Date "2020.04.30"]
[Round "1"]
[White "PlayKomodo (Komodo MCTS 32c)"]
[Black "hissha (GM Larry Kaufman"]
[Result "1-0"]
[SetUp "1"]
[FEN "rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1"]
[TimeControl "900+10"]
[WhiteElo "2960"]
[BlackElo "2267"]
[Termination "PlayKomodo won by resignation"]

1. f4 {[%clk 0:15:06.6]} 1... f5 {[%clk 0:14:54.3]} 2. e3 {[%clk 0:15:16.5]}
2... Qf7 {[%clk 0:14:34.4]} 3. b3 {[%clk 0:14:41.4]} 3... e6 {[%clk 0:14:18.7]}
4. Bf3 {[%clk 0:13:55.1]} 4... Bf6 {[%clk 0:14:19.9]} 5. d4 {[%clk 0:13:27.9]}
5... d5 {[%clk 0:14:15.4]} 6. Ba3 {[%clk 0:13:08.5]} 6... Ng6 {[%clk 0:14:02.2]}
7. Qf2 {[%clk 0:12:16.6]} 7... Be7 {[%clk 0:12:25]} 8. c4 {[%clk 0:11:19.4]}
8... Bxa3 {[%clk 0:12:17.9]} 9. Nxa3 {[%clk 0:11:00.7]} 9... c6 {[%clk
0:11:44.6]} 10. Nc2 {[%clk 0:10:22.1]} 10... O-O {[%clk 0:11:41.8]} 11. h4
{[%clk 0:09:52.7]} 11... Ne7 {[%clk 0:11:13.4]} 12. h5 {[%clk 0:09:43.7]} 12...
h6 {[%clk 0:10:52.4]} 13. O-O {[%clk 0:09:25.2]} 13... Nd7 {[%clk 0:10:47.6]}
14. a4 {[%clk 0:08:51.9]} 14... Nf6 {[%clk 0:10:38.9]} 15. a5 {[%clk 0:08:25.7]}
15... Nxh5 {[%clk 0:08:53.5]} 16. Nb4 {[%clk 0:08:35.6]} 16... a6 {[%clk
0:08:26.3]} 17. Nd3 {[%clk 0:08:10.1]} 17... Nf6 {[%clk 0:08:23.6]} 18. Ne5
{[%clk 0:07:45.6]} 18... Qe8 {[%clk 0:08:24.9]} 19. c5 {[%clk 0:07:18.1]} 19...
Ng6 {[%clk 0:08:07.4]} 20. Nd3 {[%clk 0:06:49.5]} 20... Nh8 {[%clk 0:07:44.5]}
21. g3 {[%clk 0:06:07.1]} 21... Nf7 {[%clk 0:07:40]} 22. Qe1 {[%clk 0:05:23.3]}
22... Qe7 {[%clk 0:07:00.8]} 23. Kf2 {[%clk 0:04:51.1]} 23... g5 {[%clk
0:06:40.1]} 24. Rh1 {[%clk 0:04:23.3]} 24... Kg7 {[%clk 0:06:19.2]} 25. Ke2
{[%clk 0:04:30.6]} 25... Bd7 {[%clk 0:04:35.2]} 26. Kd1 {[%clk 0:03:56.4]} 26...
Rh8 {[%clk 0:04:35.8]} 27. Kc1 {[%clk 0:03:33.6]} 27... Rag8 {[%clk 0:04:34.6]}
28. Bd1 {[%clk 0:03:00.7]} 28... Bc8 {[%clk 0:04:24.5]} 29. Ra2 {[%clk
0:02:35.4]} 29... Kf8 {[%clk 0:04:21.8]} 30. Rah2 {[%clk 0:02:13.2]} 30... Ne4
{[%clk 0:03:37.9]} 31. Kb2 {[%clk 0:01:50.3]} 31... gxf4 {[%clk 0:03:10]} 32.
gxf4 {[%clk 0:01:46]} 32... Rg3 {[%clk 0:02:38.6]} 33. Bh5 {[%clk 0:01:32.7]}
33... Rhg8 {[%clk 0:02:20.2]} 34. Bxf7 {[%clk 0:01:13.4]} 34... Qxf7 {[%clk
0:02:26.3]} 35. Rxh6 {[%clk 0:01:07.2]} 35... Rg2+ {[%clk 0:02:18.5]} 36. Ka3
{[%clk 0:01:05.1]} 36... Rd2 {[%clk 0:01:48.3]} 37. Rh7 {[%clk 0:01:00.4]} 37...
Qf6 {[%clk 0:00:49.7]} 38. Ne5 {[%clk 0:00:57.1]} 38... Qd8 {[%clk 0:00:31.4]}
39. b4 {[%clk 0:01:07]} 39... Rc2 {[%clk 0:00:17.1]} 40. Rg1 {[%clk 0:00:59.2]}
40... Rcg2 {[%clk 0:00:16.5]} 41. Rxg2 {[%clk 0:00:42.9]} 41... Rxg2 {[%clk
0:00:22.6]} 42. Qh1 {[%clk 0:00:47.2]} 42... Rg8 {[%clk 0:00:14.6]} 43. Qh6+
{[%clk 0:00:46.2]} 43... Ke8 {[%clk 0:00:22.7]} 44. Rg7 {[%clk 0:00:40.3]} 44...
Rxg7 {[%clk 0:00:15.5]} 45. Qh8+ {[%clk 0:00:41.1]} 1-0

[Chessqueen]

Looks like you picked a reasonable position and did everything properly. The final position was probably theoretically still losing for Komodo but probably too difficult for the human to win with just increment. Most of the moves looked either best or close to best on my 8 core laptop; only suspicious move was 8.c3?! but his answer was bad so it worked out ok. Did you determine whether you could use two threads or just the one? We still don't know whether MCTS actually helps in these knight (or knight for pawn) odds games vs. humans; overall results have been better, but not by enough to claim statistical significance. At least this game does suggest that the 960 factor makes the knight odds a little less easy for the human than normal knight odds.

Well I tried two threads and it was not any better but when I try 4 threads I saw an improvements up to 2200 kN/s, but I decided to use MCTS, the final position was losing for Komodo, but according to Mauricio he asked for a draw and I accepted, because at the very beginning he told me if I come close to only 20 seconds even if I am winning I will offer a draw because from experience the engine will play extremely fast without blundering but there is no way for me to beat the engine and finding the best moves with less than 20 seconds. Finally thought is that GM Lenderman will still need to remove on of his b7, c7,d7, or e7 pawn with Chess960 to make it competitive even with your super fast 32 cores system

Larry can you or your son play to us at least one game Versus your 32 core with this Chess960 Position. We would like to see how convincing your system WIN with this position using all that calculating power with the MCTS version or even a bigger surprise with your NN version, we only hear rumors but never seeing any game of your NN komodo yet. Or even better using this position pit your regular version against the MCTS or the NN version for comparison reason playing both side but without taking the knight off if it is engine vs engine , unless you use a weaker computer with black then keep this position against your 32 core monster computer just to compare:roll:

[d]rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1

Yes, I can play a game against it on my 32 core myself from that position at 15' + 10" with the MCTS version. I lose more than I win at FRC knight odds with K on this hardware at this time limit, but the most frequent result is draw. We did play our last match vs. FM Larry Gilden with the NN version, so there are games with it posted on talkchess, but it's not in proper shape to use for a test right now. Playing normal K vs. MCTS from an even position is pointless, we already have thousands of games that show that regular K is about 80 elo stronger, it doesn't much matter whether it's normal chess or 960. Playing the 32 core against a 1 core at knight odds also won't work, as we have lots of tests that show that this sort of hardware advantage only offsets something like pawn and move odds, not nearly knight odds. Normal Komodo at bullet vs. Komodo Level 23 is reasonably close at knight odds, so you could also try having your friend play normal chess (15' + 10" with Komodo level 23 (not MCTS) to see who is stronger. Komodo level 23 will move instantly, so game takes only half the time of a knight odds game with full strength komodo.

This will be interesting Can you play one game Vs the MCTS and another with the Regular komodo just to compare which will give you the hardest time ?

The way I found out how many core my system has is by holding the Ctrl + Shift + Esc and I clicked on the Performance it told me 4 core. Now if your 32 core is giving you a hard time select one with either 16 or 8 Core, since my 4 core almost got Mauricio Ruiz with this position.

I only had time for the MCTS game tonite. I could of course have easily made a draw by playing ...g4 at almost any point to lock up the game, but drawing Komodo at knight odds is no big achievement, the challenge is to win, so I went for it and lost. Maybe I can play another game without MCTS tomorrow, but it may tell us more about what I have learned from the game than whether MCTS helps or not.

[Event "Live Chess - FRCNODDS"]
[Site "Chess.com"]
[Date "2020.04.30"]
[Round "1"]
[White "PlayKomodo (Komodo MCTS 32c)"]
[Black "hissha (GM Larry Kaufman"]
[Result "1-0"]
[SetUp "1"]
[FEN "rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1"]
[TimeControl "900+10"]
[WhiteElo "2960"]
[BlackElo "2267"]
[Termination "PlayKomodo won by resignation"]

1. f4 {[%clk 0:15:06.6]} 1... f5 {[%clk 0:14:54.3]} 2. e3 {[%clk 0:15:16.5]}
2... Qf7 {[%clk 0:14:34.4]} 3. b3 {[%clk 0:14:41.4]} 3... e6 {[%clk 0:14:18.7]}
4. Bf3 {[%clk 0:13:55.1]} 4... Bf6 {[%clk 0:14:19.9]} 5. d4 {[%clk 0:13:27.9]}
5... d5 {[%clk 0:14:15.4]} 6. Ba3 {[%clk 0:13:08.5]} 6... Ng6 {[%clk 0:14:02.2]}
7. Qf2 {[%clk 0:12:16.6]} 7... Be7 {[%clk 0:12:25]} 8. c4 {[%clk 0:11:19.4]}
8... Bxa3 {[%clk 0:12:17.9]} 9. Nxa3 {[%clk 0:11:00.7]} 9... c6 {[%clk
0:11:44.6]} 10. Nc2 {[%clk 0:10:22.1]} 10... O-O {[%clk 0:11:41.8]} 11. h4
{[%clk 0:09:52.7]} 11... Ne7 {[%clk 0:11:13.4]} 12. h5 {[%clk 0:09:43.7]} 12...
h6 {[%clk 0:10:52.4]} 13. O-O {[%clk 0:09:25.2]} 13... Nd7 {[%clk 0:10:47.6]}
14. a4 {[%clk 0:08:51.9]} 14... Nf6 {[%clk 0:10:38.9]} 15. a5 {[%clk 0:08:25.7]}
15... Nxh5 {[%clk 0:08:53.5]} 16. Nb4 {[%clk 0:08:35.6]} 16... a6 {[%clk
0:08:26.3]} 17. Nd3 {[%clk 0:08:10.1]} 17... Nf6 {[%clk 0:08:23.6]} 18. Ne5
{[%clk 0:07:45.6]} 18... Qe8 {[%clk 0:08:24.9]} 19. c5 {[%clk 0:07:18.1]} 19...
Ng6 {[%clk 0:08:07.4]} 20. Nd3 {[%clk 0:06:49.5]} 20... Nh8 {[%clk 0:07:44.5]}
21. g3 {[%clk 0:06:07.1]} 21... Nf7 {[%clk 0:07:40]} 22. Qe1 {[%clk 0:05:23.3]}
22... Qe7 {[%clk 0:07:00.8]} 23. Kf2 {[%clk 0:04:51.1]} 23... g5 {[%clk
0:06:40.1]} 24. Rh1 {[%clk 0:04:23.3]} 24... Kg7 {[%clk 0:06:19.2]} 25. Ke2
{[%clk 0:04:30.6]} 25... Bd7 {[%clk 0:04:35.2]} 26. Kd1 {[%clk 0:03:56.4]} 26...
Rh8 {[%clk 0:04:35.8]} 27. Kc1 {[%clk 0:03:33.6]} 27... Rag8 {[%clk 0:04:34.6]}
28. Bd1 {[%clk 0:03:00.7]} 28... Bc8 {[%clk 0:04:24.5]} 29. Ra2 {[%clk
0:02:35.4]} 29... Kf8 {[%clk 0:04:21.8]} 30. Rah2 {[%clk 0:02:13.2]} 30... Ne4
{[%clk 0:03:37.9]} 31. Kb2 {[%clk 0:01:50.3]} 31... gxf4 {[%clk 0:03:10]} 32.
gxf4 {[%clk 0:01:46]} 32... Rg3 {[%clk 0:02:38.6]} 33. Bh5 {[%clk 0:01:32.7]}
33... Rhg8 {[%clk 0:02:20.2]} 34. Bxf7 {[%clk 0:01:13.4]} 34... Qxf7 {[%clk
0:02:26.3]} 35. Rxh6 {[%clk 0:01:07.2]} 35... Rg2+ {[%clk 0:02:18.5]} 36. Ka3
{[%clk 0:01:05.1]} 36... Rd2 {[%clk 0:01:48.3]} 37. Rh7 {[%clk 0:01:00.4]} 37...
Qf6 {[%clk 0:00:49.7]} 38. Ne5 {[%clk 0:00:57.1]} 38... Qd8 {[%clk 0:00:31.4]}
39. b4 {[%clk 0:01:07]} 39... Rc2 {[%clk 0:00:17.1]} 40. Rg1 {[%clk 0:00:59.2]}
40... Rcg2 {[%clk 0:00:16.5]} 41. Rxg2 {[%clk 0:00:42.9]} 41... Rxg2 {[%clk
0:00:22.6]} 42. Qh1 {[%clk 0:00:47.2]} 42... Rg8 {[%clk 0:00:14.6]} 43. Qh6+
{[%clk 0:00:46.2]} 43... Ke8 {[%clk 0:00:22.7]} 44. Rg7 {[%clk 0:00:40.3]} 44...
Rxg7 {[%clk 0:00:15.5]} 45. Qh8+ {[%clk 0:00:41.1]} 1-0

You did great for lasting so long against Komodo MCTS with a 32 core, If you can play the other game with either a 16 or 8 core I am pretty sure that you will defeat Komodo

You should tell GM Lenderman that a draw does NOT count for Money, since Most GM can easily close the game with pawns and only draw by three forced repetitions or stalemate count, but that you will increase by $20.00 every win, like that he will try harder and probably blunder.

[pgn]
[Event "Live Chess - FRCNODDS"]
[Site "Chess.com"]
[Date "2020.04.30"]
[Round "1"]
[White "PlayKomodo (Komodo MCTS 32c)"]
[Black "hissha (GM Larry Kaufman"]
[Result "1-0"]
[SetUp "1"]
[FEN "rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1"]
[TimeControl "900+10"]
[WhiteElo "2960"]
[BlackElo "2267"]
[Termination "PlayKomodo won by resignation"]

1. f4 {[%clk 0:15:06.6]} 1... f5 {[%clk 0:14:54.3]} 2. e3 {[%clk 0:15:16.5]}
2... Qf7 {[%clk 0:14:34.4]} 3. b3 {[%clk 0:14:41.4]} 3... e6 {[%clk 0:14:18.7]}
4. Bf3 {[%clk 0:13:55.1]} 4... Bf6 {[%clk 0:14:19.9]} 5. d4 {[%clk 0:13:27.9]}
5... d5 {[%clk 0:14:15.4]} 6. Ba3 {[%clk 0:13:08.5]} 6... Ng6 {[%clk 0:14:02.2]}
7. Qf2 {[%clk 0:12:16.6]} 7... Be7 {[%clk 0:12:25]} 8. c4 {[%clk 0:11:19.4]}
8... Bxa3 {[%clk 0:12:17.9]} 9. Nxa3 {[%clk 0:11:00.7]} 9... c6 {[%clk
0:11:44.6]} 10. Nc2 {[%clk 0:10:22.1]} 10... O-O {[%clk 0:11:41.8]} 11. h4
{[%clk 0:09:52.7]} 11... Ne7 {[%clk 0:11:13.4]} 12. h5 {[%clk 0:09:43.7]} 12...
h6 {[%clk 0:10:52.4]} 13. O-O {[%clk 0:09:25.2]} 13... Nd7 {[%clk 0:10:47.6]}
14. a4 {[%clk 0:08:51.9]} 14... Nf6 {[%clk 0:10:38.9]} 15. a5 {[%clk 0:08:25.7]}
15... Nxh5 {[%clk 0:08:53.5]} 16. Nb4 {[%clk 0:08:35.6]} 16... a6 {[%clk
0:08:26.3]} 17. Nd3 {[%clk 0:08:10.1]} 17... Nf6 {[%clk 0:08:23.6]} 18. Ne5
{[%clk 0:07:45.6]} 18... Qe8 {[%clk 0:08:24.9]} 19. c5 {[%clk 0:07:18.1]} 19...
Ng6 {[%clk 0:08:07.4]} 20. Nd3 {[%clk 0:06:49.5]} 20... Nh8 {[%clk 0:07:44.5]}
21. g3 {[%clk 0:06:07.1]} 21... Nf7 {[%clk 0:07:40]} 22. Qe1 {[%clk 0:05:23.3]}
22... Qe7 {[%clk 0:07:00.8]} 23. Kf2 {[%clk 0:04:51.1]} 23... g5 {[%clk
0:06:40.1]} 24. Rh1 {[%clk 0:04:23.3]} 24... Kg7 {[%clk 0:06:19.2]} 25. Ke2
{[%clk 0:04:30.6]} 25... Bd7 {[%clk 0:04:35.2]} 26. Kd1 {[%clk 0:03:56.4]} 26...
Rh8 {[%clk 0:04:35.8]} 27. Kc1 {[%clk 0:03:33.6]} 27... Rag8 {[%clk 0:04:34.6]}
28. Bd1 {[%clk 0:03:00.7]} 28... Bc8 {[%clk 0:04:24.5]} 29. Ra2 {[%clk
0:02:35.4]} 29... Kf8 {[%clk 0:04:21.8]} 30. Rah2 {[%clk 0:02:13.2]} 30... Ne4
{[%clk 0:03:37.9]} 31. Kb2 {[%clk 0:01:50.3]} 31... gxf4 {[%clk 0:03:10]} 32.
gxf4 {[%clk 0:01:46]} 32... Rg3 {[%clk 0:02:38.6]} 33. Bh5 {[%clk 0:01:32.7]}
33... Rhg8 {[%clk 0:02:20.2]} 34. Bxf7 {[%clk 0:01:13.4]} 34... Qxf7 {[%clk
0:02:26.3]} 35. Rxh6 {[%clk 0:01:07.2]} 35... Rg2+ {[%clk 0:02:18.5]} 36. Ka3
{[%clk 0:01:05.1]} 36... Rd2 {[%clk 0:01:48.3]} 37. Rh7 {[%clk 0:01:00.4]} 37...
Qf6 {[%clk 0:00:49.7]} 38. Ne5 {[%clk 0:00:57.1]} 38... Qd8 {[%clk 0:00:31.4]}
39. b4 {[%clk 0:01:07]} 39... Rc2 {[%clk 0:00:17.1]} 40. Rg1 {[%clk 0:00:59.2]}
40... Rcg2 {[%clk 0:00:16.5]} 41. Rxg2 {[%clk 0:00:42.9]} 41... Rxg2 {[%clk
0:00:22.6]} 42. Qh1 {[%clk 0:00:47.2]} 42... Rg8 {[%clk 0:00:14.6]} 43. Qh6+
{[%clk 0:00:46.2]} 43... Ke8 {[%clk 0:00:22.7]} 44. Rg7 {[%clk 0:00:40.3]} 44...
Rxg7 {[%clk 0:00:15.5]} 45. Qh8+ {[%clk 0:00:41.1]} 1-0[/pgn]

[lkaufman]

Looks like you picked a reasonable position and did everything properly. The final position was probably theoretically still losing for Komodo but probably too difficult for the human to win with just increment. Most of the moves looked either best or close to best on my 8 core laptop; only suspicious move was 8.c3?! but his answer was bad so it worked out ok. Did you determine whether you could use two threads or just the one? We still don't know whether MCTS actually helps in these knight (or knight for pawn) odds games vs. humans; overall results have been better, but not by enough to claim statistical significance. At least this game does suggest that the 960 factor makes the knight odds a little less easy for the human than normal knight odds.

Well I tried two threads and it was not any better but when I try 4 threads I saw an improvements up to 2200 kN/s, but I decided to use MCTS, the final position was losing for Komodo, but according to Mauricio he asked for a draw and I accepted, because at the very beginning he told me if I come close to only 20 seconds even if I am winning I will offer a draw because from experience the engine will play extremely fast without blundering but there is no way for me to beat the engine and finding the best moves with less than 20 seconds. Finally thought is that GM Lenderman will still need to remove on of his b7, c7,d7, or e7 pawn with Chess960 to make it competitive even with your super fast 32 cores system

Larry can you or your son play to us at least one game Versus your 32 core with this Chess960 Position. We would like to see how convincing your system WIN with this position using all that calculating power with the MCTS version or even a bigger surprise with your NN version, we only hear rumors but never seeing any game of your NN komodo yet. Or even better using this position pit your regular version against the MCTS or the NN version for comparison reason playing both side but without taking the knight off if it is engine vs engine , unless you use a weaker computer with black then keep this position against your 32 core monster computer just to compare:roll:

[d]rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1

Yes, I can play a game against it on my 32 core myself from that position at 15' + 10" with the MCTS version. I lose more than I win at FRC knight odds with K on this hardware at this time limit, but the most frequent result is draw. We did play our last match vs. FM Larry Gilden with the NN version, so there are games with it posted on talkchess, but it's not in proper shape to use for a test right now. Playing normal K vs. MCTS from an even position is pointless, we already have thousands of games that show that regular K is about 80 elo stronger, it doesn't much matter whether it's normal chess or 960. Playing the 32 core against a 1 core at knight odds also won't work, as we have lots of tests that show that this sort of hardware advantage only offsets something like pawn and move odds, not nearly knight odds. Normal Komodo at bullet vs. Komodo Level 23 is reasonably close at knight odds, so you could also try having your friend play normal chess (15' + 10" with Komodo level 23 (not MCTS) to see who is stronger. Komodo level 23 will move instantly, so game takes only half the time of a knight odds game with full strength komodo.

This will be interesting Can you play one game Vs the MCTS and another with the Regular komodo just to compare which will give you the hardest time ?

The way I found out how many core my system has is by holding the Ctrl + Shift + Esc and I clicked on the Performance it told me 4 core. Now if your 32 core is giving you a hard time select one with either 16 or 8 Core, since my 4 core almost got Mauricio Ruiz with this position.

I only had time for the MCTS game tonite. I could of course have easily made a draw by playing ...g4 at almost any point to lock up the game, but drawing Komodo at knight odds is no big achievement, the challenge is to win, so I went for it and lost. Maybe I can play another game without MCTS tomorrow, but it may tell us more about what I have learned from the game than whether MCTS helps or not.

[Event "Live Chess - FRCNODDS"]
[Site "Chess.com"]
[Date "2020.04.30"]
[Round "1"]
[White "PlayKomodo (Komodo MCTS 32c)"]
[Black "hissha (GM Larry Kaufman"]
[Result "1-0"]
[SetUp "1"]
[FEN "rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1"]
[TimeControl "900+10"]
[WhiteElo "2960"]
[BlackElo "2267"]
[Termination "PlayKomodo won by resignation"]

1. f4 {[%clk 0:15:06.6]} 1... f5 {[%clk 0:14:54.3]} 2. e3 {[%clk 0:15:16.5]}
2... Qf7 {[%clk 0:14:34.4]} 3. b3 {[%clk 0:14:41.4]} 3... e6 {[%clk 0:14:18.7]}
4. Bf3 {[%clk 0:13:55.1]} 4... Bf6 {[%clk 0:14:19.9]} 5. d4 {[%clk 0:13:27.9]}
5... d5 {[%clk 0:14:15.4]} 6. Ba3 {[%clk 0:13:08.5]} 6... Ng6 {[%clk 0:14:02.2]}
7. Qf2 {[%clk 0:12:16.6]} 7... Be7 {[%clk 0:12:25]} 8. c4 {[%clk 0:11:19.4]}
8... Bxa3 {[%clk 0:12:17.9]} 9. Nxa3 {[%clk 0:11:00.7]} 9... c6 {[%clk
0:11:44.6]} 10. Nc2 {[%clk 0:10:22.1]} 10... O-O {[%clk 0:11:41.8]} 11. h4
{[%clk 0:09:52.7]} 11... Ne7 {[%clk 0:11:13.4]} 12. h5 {[%clk 0:09:43.7]} 12...
h6 {[%clk 0:10:52.4]} 13. O-O {[%clk 0:09:25.2]} 13... Nd7 {[%clk 0:10:47.6]}
14. a4 {[%clk 0:08:51.9]} 14... Nf6 {[%clk 0:10:38.9]} 15. a5 {[%clk 0:08:25.7]}
15... Nxh5 {[%clk 0:08:53.5]} 16. Nb4 {[%clk 0:08:35.6]} 16... a6 {[%clk
0:08:26.3]} 17. Nd3 {[%clk 0:08:10.1]} 17... Nf6 {[%clk 0:08:23.6]} 18. Ne5
{[%clk 0:07:45.6]} 18... Qe8 {[%clk 0:08:24.9]} 19. c5 {[%clk 0:07:18.1]} 19...
Ng6 {[%clk 0:08:07.4]} 20. Nd3 {[%clk 0:06:49.5]} 20... Nh8 {[%clk 0:07:44.5]}
21. g3 {[%clk 0:06:07.1]} 21... Nf7 {[%clk 0:07:40]} 22. Qe1 {[%clk 0:05:23.3]}
22... Qe7 {[%clk 0:07:00.8]} 23. Kf2 {[%clk 0:04:51.1]} 23... g5 {[%clk
0:06:40.1]} 24. Rh1 {[%clk 0:04:23.3]} 24... Kg7 {[%clk 0:06:19.2]} 25. Ke2
{[%clk 0:04:30.6]} 25... Bd7 {[%clk 0:04:35.2]} 26. Kd1 {[%clk 0:03:56.4]} 26...
Rh8 {[%clk 0:04:35.8]} 27. Kc1 {[%clk 0:03:33.6]} 27... Rag8 {[%clk 0:04:34.6]}
28. Bd1 {[%clk 0:03:00.7]} 28... Bc8 {[%clk 0:04:24.5]} 29. Ra2 {[%clk
0:02:35.4]} 29... Kf8 {[%clk 0:04:21.8]} 30. Rah2 {[%clk 0:02:13.2]} 30... Ne4
{[%clk 0:03:37.9]} 31. Kb2 {[%clk 0:01:50.3]} 31... gxf4 {[%clk 0:03:10]} 32.
gxf4 {[%clk 0:01:46]} 32... Rg3 {[%clk 0:02:38.6]} 33. Bh5 {[%clk 0:01:32.7]}
33... Rhg8 {[%clk 0:02:20.2]} 34. Bxf7 {[%clk 0:01:13.4]} 34... Qxf7 {[%clk
0:02:26.3]} 35. Rxh6 {[%clk 0:01:07.2]} 35... Rg2+ {[%clk 0:02:18.5]} 36. Ka3
{[%clk 0:01:05.1]} 36... Rd2 {[%clk 0:01:48.3]} 37. Rh7 {[%clk 0:01:00.4]} 37...
Qf6 {[%clk 0:00:49.7]} 38. Ne5 {[%clk 0:00:57.1]} 38... Qd8 {[%clk 0:00:31.4]}
39. b4 {[%clk 0:01:07]} 39... Rc2 {[%clk 0:00:17.1]} 40. Rg1 {[%clk 0:00:59.2]}
40... Rcg2 {[%clk 0:00:16.5]} 41. Rxg2 {[%clk 0:00:42.9]} 41... Rxg2 {[%clk
0:00:22.6]} 42. Qh1 {[%clk 0:00:47.2]} 42... Rg8 {[%clk 0:00:14.6]} 43. Qh6+
{[%clk 0:00:46.2]} 43... Ke8 {[%clk 0:00:22.7]} 44. Rg7 {[%clk 0:00:40.3]} 44...
Rxg7 {[%clk 0:00:15.5]} 45. Qh8+ {[%clk 0:00:41.1]} 1-0

You did great for lasting so long against Komodo MCTS with a 32 core, If you can play the other game with either a 16 or 8 core I am pretty sure that you will defeat Komodo

You should tell GM Lenderman that a draw does NOT count for Money, since Most GM can easily close the game with pawns and only draw by three forced repetitions or stalemate count, but that you will increase by $20.00 every win, like that he will try harder and probably blunder.

[pgn]
[Event "Live Chess - FRCNODDS"]
[Site "Chess.com"]
[Date "2020.04.30"]
[Round "1"]
[White "PlayKomodo (Komodo MCTS 32c)"]
[Black "hissha (GM Larry Kaufman"]
[Result "1-0"]
[SetUp "1"]
[FEN "rnbbknqr/pppppppp/8/8/8/8/PPPPPPPP/RNBBK1QR w KQkq - 0 1"]
[TimeControl "900+10"]
[WhiteElo "2960"]
[BlackElo "2267"]
[Termination "PlayKomodo won by resignation"]

1. f4 {[%clk 0:15:06.6]} 1... f5 {[%clk 0:14:54.3]} 2. e3 {[%clk 0:15:16.5]}
2... Qf7 {[%clk 0:14:34.4]} 3. b3 {[%clk 0:14:41.4]} 3... e6 {[%clk 0:14:18.7]}
4. Bf3 {[%clk 0:13:55.1]} 4... Bf6 {[%clk 0:14:19.9]} 5. d4 {[%clk 0:13:27.9]}
5... d5 {[%clk 0:14:15.4]} 6. Ba3 {[%clk 0:13:08.5]} 6... Ng6 {[%clk 0:14:02.2]}
7. Qf2 {[%clk 0:12:16.6]} 7... Be7 {[%clk 0:12:25]} 8. c4 {[%clk 0:11:19.4]}
8... Bxa3 {[%clk 0:12:17.9]} 9. Nxa3 {[%clk 0:11:00.7]} 9... c6 {[%clk
0:11:44.6]} 10. Nc2 {[%clk 0:10:22.1]} 10... O-O {[%clk 0:11:41.8]} 11. h4
{[%clk 0:09:52.7]} 11... Ne7 {[%clk 0:11:13.4]} 12. h5 {[%clk 0:09:43.7]} 12...
h6 {[%clk 0:10:52.4]} 13. O-O {[%clk 0:09:25.2]} 13... Nd7 {[%clk 0:10:47.6]}
14. a4 {[%clk 0:08:51.9]} 14... Nf6 {[%clk 0:10:38.9]} 15. a5 {[%clk 0:08:25.7]}
15... Nxh5 {[%clk 0:08:53.5]} 16. Nb4 {[%clk 0:08:35.6]} 16... a6 {[%clk
0:08:26.3]} 17. Nd3 {[%clk 0:08:10.1]} 17... Nf6 {[%clk 0:08:23.6]} 18. Ne5
{[%clk 0:07:45.6]} 18... Qe8 {[%clk 0:08:24.9]} 19. c5 {[%clk 0:07:18.1]} 19...
Ng6 {[%clk 0:08:07.4]} 20. Nd3 {[%clk 0:06:49.5]} 20... Nh8 {[%clk 0:07:44.5]}
21. g3 {[%clk 0:06:07.1]} 21... Nf7 {[%clk 0:07:40]} 22. Qe1 {[%clk 0:05:23.3]}
22... Qe7 {[%clk 0:07:00.8]} 23. Kf2 {[%clk 0:04:51.1]} 23... g5 {[%clk
0:06:40.1]} 24. Rh1 {[%clk 0:04:23.3]} 24... Kg7 {[%clk 0:06:19.2]} 25. Ke2
{[%clk 0:04:30.6]} 25... Bd7 {[%clk 0:04:35.2]} 26. Kd1 {[%clk 0:03:56.4]} 26...
Rh8 {[%clk 0:04:35.8]} 27. Kc1 {[%clk 0:03:33.6]} 27... Rag8 {[%clk 0:04:34.6]}
28. Bd1 {[%clk 0:03:00.7]} 28... Bc8 {[%clk 0:04:24.5]} 29. Ra2 {[%clk
0:02:35.4]} 29... Kf8 {[%clk 0:04:21.8]} 30. Rah2 {[%clk 0:02:13.2]} 30... Ne4
{[%clk 0:03:37.9]} 31. Kb2 {[%clk 0:01:50.3]} 31... gxf4 {[%clk 0:03:10]} 32.
gxf4 {[%clk 0:01:46]} 32... Rg3 {[%clk 0:02:38.6]} 33. Bh5 {[%clk 0:01:32.7]}
33... Rhg8 {[%clk 0:02:20.2]} 34. Bxf7 {[%clk 0:01:13.4]} 34... Qxf7 {[%clk
0:02:26.3]} 35. Rxh6 {[%clk 0:01:07.2]} 35... Rg2+ {[%clk 0:02:18.5]} 36. Ka3
{[%clk 0:01:05.1]} 36... Rd2 {[%clk 0:01:48.3]} 37. Rh7 {[%clk 0:01:00.4]} 37...
Qf6 {[%clk 0:00:49.7]} 38. Ne5 {[%clk 0:00:57.1]} 38... Qd8 {[%clk 0:00:31.4]}
39. b4 {[%clk 0:01:07]} 39... Rc2 {[%clk 0:00:17.1]} 40. Rg1 {[%clk 0:00:59.2]}
40... Rcg2 {[%clk 0:00:16.5]} 41. Rxg2 {[%clk 0:00:42.9]} 41... Rxg2 {[%clk
0:00:22.6]} 42. Qh1 {[%clk 0:00:47.2]} 42... Rg8 {[%clk 0:00:14.6]} 43. Qh6+
{[%clk 0:00:46.2]} 43... Ke8 {[%clk 0:00:22.7]} 44. Rg7 {[%clk 0:00:40.3]} 44...
Rxg7 {[%clk 0:00:15.5]} 45. Qh8+ {[%clk 0:00:41.1]} 1-0[/pgn]

Lenderman will get only $10 prize money for draws but $30 for wins. He definitely won't aim for draws at these handicaps, but he does deserve some reward for saving a draw once he loses his advantage.