Amazing results of LeelaKnightOdds-Dev

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

Moderator: Ras

lkaufman
Posts: 6147
Joined: Sun Jan 10, 2010 6:15 am
Location: Maryland USA
Full name: Larry Kaufman

Re: Amazing results of LeelaKnightOdds-Dev

Post by lkaufman »

Father wrote: Wed Jan 15, 2025 2:11 pm
... apparently the robots are asleep or resting, my friend LeelaQueenOdds doesn't call me, she doesn't answer my messages, it seems like she's stopped wanting to dance with me. I'm on the bench alone and despised.
The problem has been fixed. The bots were transferred to a different computer, which was missing one needed file. Should be identical to yesterday's bots now.
Komodo rules!
Father
Posts: 1722
Joined: Sun Mar 19, 2006 4:39 am
Location: Colombia
Full name: Pablo Ignacio Restrepo

Re: Amazing results of LeelaKnightOdds-Dev

Post by Father »

lkaufman wrote: Wed Jan 15, 2025 5:15 pm
Father wrote: Wed Jan 15, 2025 2:11 pm
... apparently the robots are asleep or resting, my friend LeelaQueenOdds doesn't call me, she doesn't answer my messages, it seems like she's stopped wanting to dance with me. I'm on the bench alone and despised.
The problem has been fixed. The bots were transferred to a different computer, which was missing one needed file. Should be identical to yesterday's bots now.
It could happen how things go that the prize for humans facing
LeelaQueenOdds is to occupy a place in the table of the 100 best humans. It turns out that the table of the 100 best is equivalent to a big three bagon, where there are only 100 berths for 100 people, but the number of people who want to travel on the train is substantially higher than the quota. Long live LeelaQueenOdds.!!
I am thinking chess is in a coin.Human beings for ever playing in one face.Now I am playing in the other face:"Antichess". Computers are as a fortres where owner forgot to close a little door behind. You must enter across this door.Forget the front.
Father
Posts: 1722
Joined: Sun Mar 19, 2006 4:39 am
Location: Colombia
Full name: Pablo Ignacio Restrepo

Re: Amazing results of LeelaKnightOdds-Dev

Post by Father »

Father wrote: Wed Jan 15, 2025 11:25 pm
lkaufman wrote: Wed Jan 15, 2025 5:15 pm
Father wrote: Wed Jan 15, 2025 2:11 pm
... apparently the robots are asleep or resting, my friend LeelaQueenOdds doesn't call me, she doesn't answer my messages, it seems like she's stopped wanting to dance with me. I'm on the bench alone and despised.
The problem has been fixed. The bots were transferred to a different computer, which was missing one needed file. Should be identical to yesterday's bots now.
It could happen how things go that the prize for humans facing
LeelaQueenOdds is to occupy a place in the table of the 100 best humans. It turns out that the table of the 100 best is equivalent to a big three bagon, where there are only 100 berths for 100 people, but the number of people who want to travel on the train is substantially higher than the quota. Long live LeelaQueenOdds.!!
I wish to remember one particular game between man and machine LeelaQueenOdds. I was just watching this game, which, in my own understanding of chess, shows me the greatness that is a good draw.

The human Hissha played fearlessly, and not even the rush of time affected tactical moves in the final phase of the game, like a trapeze artist with nerves of steel. For its part, the machine made use of extraordinary resources. Such moves are worth much more than many victories. How great is the human being who invented chess, machines and robots. Life is beautiful, life is life.

[pgn][Event "Casual blitz game"]
[Site "https://lichess.org/xOR1l3OS"]
[Date "2025.01.16"]
[White "LeelaQueenOdds"]
[Black "Hissha"]
[Result "1/2-1/2"]
[UTCDate "2025.01.16"]
[UTCTime "04:02:57"]
[WhiteElo "2000"]
[BlackElo "2374"]
[WhiteTitle "BOT"]
[Variant "From Position"]
[TimeControl "60+10"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNB1KBNR w KQkq - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. g3 d5 2. Nf3 Bf5 3. d3 Nf6 4. Nc3 e6 5. h3 d4 6. Nb1 Nc6 7. g4 Bg6 8. a3 Bd6 9. Nbd2 e5 10. g5 Nd5 11. h4 Bh5 12. Ne4 Bxf3 13. exf3 Qe7 14. Bh3 Nf4 15. Bxf4 exf4 16. O-O-O Ne5 17. Rde1 O-O 18. Bf5 Nxf3 19. Ref1 Ne5 20. Rfg1 Bxa3 21. h5 Bb4 22. Rg2 a5 23. Kb1 a4 24. Ka2 a3 25. b3 b5 26. f3 c5 27. Nf6+ Kh8 28. Nxh7 c4 29. g6 cxb3+ 30. cxb3 Rfc8 31. Be4 Ra7 32. Ng5 Rac7 33. Rhh2 Rc1 34. h6 fxg6 35. hxg7+ Kxg7 36. Bd5 Rh8 37. Rxh8 Kxh8 38. Rh2+ Kg7 39. Rh7+ Kf8 40. Rh8+ Kg7 41. Rh7+ Kf8 42. Rh8+ Kg7 43. Rh7+ { The game is a draw. } 1/2-1/2[/pgn]
I am thinking chess is in a coin.Human beings for ever playing in one face.Now I am playing in the other face:"Antichess". Computers are as a fortres where owner forgot to close a little door behind. You must enter across this door.Forget the front.
Father
Posts: 1722
Joined: Sun Mar 19, 2006 4:39 am
Location: Colombia
Full name: Pablo Ignacio Restrepo

Re: Amazing results of LeelaKnightOdds-Dev

Post by Father »

Father wrote: Thu Jan 16, 2025 5:24 am
Father wrote: Wed Jan 15, 2025 11:25 pm
lkaufman wrote: Wed Jan 15, 2025 5:15 pm
Father wrote: Wed Jan 15, 2025 2:11 pm
... apparently the robots are asleep or resting, my friend LeelaQueenOdds doesn't call me, she doesn't answer my messages, it seems like she's stopped wanting to dance with me. I'm on the bench alone and despised.
The problem has been fixed. The bots were transferred to a different computer, which was missing one needed file. Should be identical to yesterday's bots now.
It could happen how things go that the prize for humans facing
LeelaQueenOdds is to occupy a place in the table of the 100 best humans. It turns out that the table of the 100 best is equivalent to a big three bagon, where there are only 100 berths for 100 people, but the number of people who want to travel on the train is substantially higher than the quota. Long live LeelaQueenOdds.!!
I wish to remember one particular game between man and machine LeelaQueenOdds. I was just watching this game, which, in my own understanding of chess, shows me the greatness that is a good draw.

The human Hissha played fearlessly, and not even the rush of time affected tactical moves in the final phase of the game, like a trapeze artist with nerves of steel. For its part, the machine made use of extraordinary resources. Such moves are worth much more than many victories. How great is the human being who invented chess, machines and robots. Life is beautiful, life is life.

[pgn][Event "Casual blitz game"]
[Site "https://lichess.org/xOR1l3OS"]
[Date "2025.01.16"]
[White "LeelaQueenOdds"]
[Black "Hissha"]
[Result "1/2-1/2"]
[UTCDate "2025.01.16"]
[UTCTime "04:02:57"]
[WhiteElo "2000"]
[BlackElo "2374"]
[WhiteTitle "BOT"]
[Variant "From Position"]
[TimeControl "60+10"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNB1KBNR w KQkq - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. g3 d5 2. Nf3 Bf5 3. d3 Nf6 4. Nc3 e6 5. h3 d4 6. Nb1 Nc6 7. g4 Bg6 8. a3 Bd6 9. Nbd2 e5 10. g5 Nd5 11. h4 Bh5 12. Ne4 Bxf3 13. exf3 Qe7 14. Bh3 Nf4 15. Bxf4 exf4 16. O-O-O Ne5 17. Rde1 O-O 18. Bf5 Nxf3 19. Ref1 Ne5 20. Rfg1 Bxa3 21. h5 Bb4 22. Rg2 a5 23. Kb1 a4 24. Ka2 a3 25. b3 b5 26. f3 c5 27. Nf6+ Kh8 28. Nxh7 c4 29. g6 cxb3+ 30. cxb3 Rfc8 31. Be4 Ra7 32. Ng5 Rac7 33. Rhh2 Rc1 34. h6 fxg6 35. hxg7+ Kxg7 36. Bd5 Rh8 37. Rxh8 Kxh8 38. Rh2+ Kg7 39. Rh7+ Kf8 40. Rh8+ Kg7 41. Rh7+ Kf8 42. Rh8+ Kg7 43. Rh7+ { The game is a draw. } 1/2-1/2[/pgn]
Good morning Mr. Larry Kaufman. I have seen that there is another super robot. "LeelaPieceOdds" I wonder... is it possible to play against that robot? The robot represents different kinds of Odds. How wonderful!! Thanks to the entire Leela team for allowing us the pleasure of playing against robots.🐴
I am thinking chess is in a coin.Human beings for ever playing in one face.Now I am playing in the other face:"Antichess". Computers are as a fortres where owner forgot to close a little door behind. You must enter across this door.Forget the front.
Uri Blass
Posts: 10704
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Amazing results of LeelaKnightOdds-Dev

Post by Uri Blass »

Father wrote: Thu Jan 16, 2025 5:24 am
Father wrote: Wed Jan 15, 2025 11:25 pm
lkaufman wrote: Wed Jan 15, 2025 5:15 pm
Father wrote: Wed Jan 15, 2025 2:11 pm
... apparently the robots are asleep or resting, my friend LeelaQueenOdds doesn't call me, she doesn't answer my messages, it seems like she's stopped wanting to dance with me. I'm on the bench alone and despised.
The problem has been fixed. The bots were transferred to a different computer, which was missing one needed file. Should be identical to yesterday's bots now.
It could happen how things go that the prize for humans facing
LeelaQueenOdds is to occupy a place in the table of the 100 best humans. It turns out that the table of the 100 best is equivalent to a big three bagon, where there are only 100 berths for 100 people, but the number of people who want to travel on the train is substantially higher than the quota. Long live LeelaQueenOdds.!!
I wish to remember one particular game between man and machine LeelaQueenOdds. I was just watching this game, which, in my own understanding of chess, shows me the greatness that is a good draw.

The human Hissha played fearlessly, and not even the rush of time affected tactical moves in the final phase of the game, like a trapeze artist with nerves of steel. For its part, the machine made use of extraordinary resources. Such moves are worth much more than many victories. How great is the human being who invented chess, machines and robots. Life is beautiful, life is life.

[pgn][Event "Casual blitz game"]
[Site "https://lichess.org/xOR1l3OS"]
[Date "2025.01.16"]
[White "LeelaQueenOdds"]
[Black "Hissha"]
[Result "1/2-1/2"]
[UTCDate "2025.01.16"]
[UTCTime "04:02:57"]
[WhiteElo "2000"]
[BlackElo "2374"]
[WhiteTitle "BOT"]
[Variant "From Position"]
[TimeControl "60+10"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNB1KBNR w KQkq - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. g3 d5 2. Nf3 Bf5 3. d3 Nf6 4. Nc3 e6 5. h3 d4 6. Nb1 Nc6 7. g4 Bg6 8. a3 Bd6 9. Nbd2 e5 10. g5 Nd5 11. h4 Bh5 12. Ne4 Bxf3 13. exf3 Qe7 14. Bh3 Nf4 15. Bxf4 exf4 16. O-O-O Ne5 17. Rde1 O-O 18. Bf5 Nxf3 19. Ref1 Ne5 20. Rfg1 Bxa3 21. h5 Bb4 22. Rg2 a5 23. Kb1 a4 24. Ka2 a3 25. b3 b5 26. f3 c5 27. Nf6+ Kh8 28. Nxh7 c4 29. g6 cxb3+ 30. cxb3 Rfc8 31. Be4 Ra7 32. Ng5 Rac7 33. Rhh2 Rc1 34. h6 fxg6 35. hxg7+ Kxg7 36. Bd5 Rh8 37. Rxh8 Kxh8 38. Rh2+ Kg7 39. Rh7+ Kf8 40. Rh8+ Kg7 41. Rh7+ Kf8 42. Rh8+ Kg7 43. Rh7+ { The game is a draw. } 1/2-1/2[/pgn]

Looking at this game I guess the human blundered because of time trouble.
41...Kf6 was winning and maybe black was afraid from 42.Ne4+ Kf5 but black is winning after Rxe7 Bxe7(it is easy to miss the fact that the rook is defended by the bishop when you do not have a lot of time and I guess Larry missed it and did not play Kf6 because of it but I am not sure if I am right and maybe it is not the reason for Kf8).

Note that I also did a similiar mistake of missing a capture in my calculation from a winning position and lost partly because of that.
[d]r3r1k1/pp3ppp/2pb4/3p3q/6nP/1PNPP1P1/PBR2P2/3Q1RK1 b - h3 0 18
Here my calculation in the game started.

I caluclated g5 Kg2 gxh4 Rh1 h3+ and thought wrongly that I win if Rxh3 Nxe3+ fxe3 Qxd1 so I thought white needs to reply Kg1 and did not pay much attention to Rxh3(I was not 100% sure if I win easily after Kg1 so played h3+ too fast to save time for me for later moves because I assumed I may need some time after h3+ Kg1.

Only after Rxh3 I understood that my opponent will not do a stupid mistake of losing the queen so I played Qg6 that is not losing but I lost later.

Here the position that I blundered because Qg6 at that point can win the game(if Rxh4 then Nxe3+)
[d]r3r1k1/pp3p1p/2pb4/3p3q/6np/1PNPP1P1/PBR2PK1/3Q3R b - - 1 20

If I try to compare the mistakes then
Note that Larry needed 5 plies to calculate Kf6 Ne4+ Kf5 Qxe7 Bxe7 when I needed to calculate in advance 6 plies to calculate in advance
h3+ Rxh3 Nxe3+ fxe3 Qxd1 Nxd1

I consider calculating 6 plies forward in advance in 90+30 to be easier than calculating 5 plies forward in advance in 1+10 but of course larry is a better player than me and I do not know if I am right by guessing the reason for his mistake.
lkaufman
Posts: 6147
Joined: Sun Jan 10, 2010 6:15 am
Location: Maryland USA
Full name: Larry Kaufman

Re: Amazing results of LeelaKnightOdds-Dev

Post by lkaufman »

Uri Blass wrote: Fri Jan 17, 2025 2:23 pm
Father wrote: Thu Jan 16, 2025 5:24 am
Father wrote: Wed Jan 15, 2025 11:25 pm
lkaufman wrote: Wed Jan 15, 2025 5:15 pm
Father wrote: Wed Jan 15, 2025 2:11 pm
... apparently the robots are asleep or resting, my friend LeelaQueenOdds doesn't call me, she doesn't answer my messages, it seems like she's stopped wanting to dance with me. I'm on the bench alone and despised.
The problem has been fixed. The bots were transferred to a different computer, which was missing one needed file. Should be identical to yesterday's bots now.
It could happen how things go that the prize for humans facing
LeelaQueenOdds is to occupy a place in the table of the 100 best humans. It turns out that the table of the 100 best is equivalent to a big three bagon, where there are only 100 berths for 100 people, but the number of people who want to travel on the train is substantially higher than the quota. Long live LeelaQueenOdds.!!
I wish to remember one particular game between man and machine LeelaQueenOdds. I was just watching this game, which, in my own understanding of chess, shows me the greatness that is a good draw.

The human Hissha played fearlessly, and not even the rush of time affected tactical moves in the final phase of the game, like a trapeze artist with nerves of steel. For its part, the machine made use of extraordinary resources. Such moves are worth much more than many victories. How great is the human being who invented chess, machines and robots. Life is beautiful, life is life.

[pgn][Event "Casual blitz game"]
[Site "https://lichess.org/xOR1l3OS"]
[Date "2025.01.16"]
[White "LeelaQueenOdds"]
[Black "Hissha"]
[Result "1/2-1/2"]
[UTCDate "2025.01.16"]
[UTCTime "04:02:57"]
[WhiteElo "2000"]
[BlackElo "2374"]
[WhiteTitle "BOT"]
[Variant "From Position"]
[TimeControl "60+10"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNB1KBNR w KQkq - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. g3 d5 2. Nf3 Bf5 3. d3 Nf6 4. Nc3 e6 5. h3 d4 6. Nb1 Nc6 7. g4 Bg6 8. a3 Bd6 9. Nbd2 e5 10. g5 Nd5 11. h4 Bh5 12. Ne4 Bxf3 13. exf3 Qe7 14. Bh3 Nf4 15. Bxf4 exf4 16. O-O-O Ne5 17. Rde1 O-O 18. Bf5 Nxf3 19. Ref1 Ne5 20. Rfg1 Bxa3 21. h5 Bb4 22. Rg2 a5 23. Kb1 a4 24. Ka2 a3 25. b3 b5 26. f3 c5 27. Nf6+ Kh8 28. Nxh7 c4 29. g6 cxb3+ 30. cxb3 Rfc8 31. Be4 Ra7 32. Ng5 Rac7 33. Rhh2 Rc1 34. h6 fxg6 35. hxg7+ Kxg7 36. Bd5 Rh8 37. Rxh8 Kxh8 38. Rh2+ Kg7 39. Rh7+ Kf8 40. Rh8+ Kg7 41. Rh7+ Kf8 42. Rh8+ Kg7 43. Rh7+ { The game is a draw. } 1/2-1/2[/pgn]

Looking at this game I guess the human blundered because of time trouble.
41...Kf6 was winning and maybe black was afraid from 42.Ne4+ Kf5 but black is winning after Rxe7 Bxe7(it is easy to miss the fact that the rook is defended by the bishop when you do not have a lot of time and I guess Larry missed it and did not play Kf6 because of it but I am not sure if I am right and maybe it is not the reason for Kf8).

Note that I also did a similiar mistake of missing a capture in my calculation from a winning position and lost partly because of that.
[d]r3r1k1/pp3ppp/2pb4/3p3q/6nP/1PNPP1P1/PBR2P2/3Q1RK1 b - h3 0 18
Here my calculation in the game started.

I caluclated g5 Kg2 gxh4 Rh1 h3+ and thought wrongly that I win if Rxh3 Nxe3+ fxe3 Qxd1 so I thought white needs to reply Kg1 and did not pay much attention to Rxh3(I was not 100% sure if I win easily after Kg1 so played h3+ too fast to save time for me for later moves because I assumed I may need some time after h3+ Kg1.

Only after Rxh3 I understood that my opponent will not do a stupid mistake of losing the queen so I played Qg6 that is not losing but I lost later.

Here the position that I blundered because Qg6 at that point can win the game(if Rxh4 then Nxe3+)
[d]r3r1k1/pp3p1p/2pb4/3p3q/6np/1PNPP1P1/PBR2PK1/3Q3R b - - 1 20

If I try to compare the mistakes then
Note that Larry needed 5 plies to calculate Kf6 Ne4+ Kf5 Qxe7 Bxe7 when I needed to calculate in advance 6 plies to calculate in advance
h3+ Rxh3 Nxe3+ fxe3 Qxd1 Nxd1

I consider calculating 6 plies forward in advance in 90+30 to be easier than calculating 5 plies forward in advance in 1+10 but of course larry is a better player than me and I do not know if I am right by guessing the reason for his mistake.
Yes, since I was playing basically on just the ten second increment, and had to split that ten seconds between considering ...Kf8 and Kf6, I believe I didn't notice that the bishop defended the queen backwards. In general, LeelaQueenOdds counts on the human missing such fairly simple tactics, otherwise it would just play to postpone defeat and would score much worse. For any given human opponent above perhaps 2000 LIchess blitz level, there is some available time limit above which the human will see enough such tactics to make a plus score, but that time limit ranges from 1'+1" for the world's best players to near the maximum 30'+30" at around 2000 level. For me, as long as (base+inc) is at least close to ten or more, I can do it. Maybe with enough practice I could reduce that to 7 or 8.
Komodo rules!
lkaufman
Posts: 6147
Joined: Sun Jan 10, 2010 6:15 am
Location: Maryland USA
Full name: Larry Kaufman

Re: Amazing results of LeelaKnightOdds-Dev

Post by lkaufman »

Father wrote: Fri Jan 17, 2025 1:10 pm
Good morning Mr. Larry Kaufman. I have seen that there is another super robot. "LeelaPieceOdds" I wonder... is it possible to play against that robot? The robot represents different kinds of Odds. How wonderful!! Thanks to the entire Leela team for allowing us the pleasure of playing against robots.🐴
I posted about this when it first came out a month or so ago. Sure, anyone with a LiChess account can play against "LeelaPieceOdds" at any of the available odds. Note that this bot runs on a computer in Europe that is not available during the night in Europe, and also that it uses the queen odds net, so it is not quite as good at playing the various odds as the queen odds bot is at playing queen odds. But it isn't a lot worse, roughly a pawn or so worse I would say. It is mostly useful for two groups of players: 1. Strong players who can generally beat Leela Queen Odds in Rapid but who have hardly any chance with just Rook odds. I am in that group, and find that Two Knights handicap is the most appropriate for me; I can generally win in Rapid but rarely in blitz, even slow blitz. Of course we already have a Queen for Knight bot that is also suitable, but it is still quite challenging even at my level; the two knights bot is easier to defeat. 2. Players below the level needed to beat LeelaQueenOdds at the time limit they like. They can choose Queen and knight, or four minors, or Queen and Rook, or Queen and two knights or bishops, or even Queen and Two rooks! Of course, for players below 1000 or so even that won't be enough to win at Rapid, and of course as the time limit gets shorter the needed rating to win goes higher. At one minute bullet chess it might take a bullet rating over 2000 to win at queen and two rooks, there isn't much data yet. I probably couldn't do it as my own level at pure one minute bullet is far below 2000, I can't play more than about 30 moves in a minute without increment.
Komodo rules!
Father
Posts: 1722
Joined: Sun Mar 19, 2006 4:39 am
Location: Colombia
Full name: Pablo Ignacio Restrepo

Re: Amazing results of LeelaKnightOdds-Dev

Post by Father »

lkaufman wrote: Fri Jan 17, 2025 7:10 pm
Father wrote: Fri Jan 17, 2025 1:10 pm
Good morning Mr. Larry Kaufman. I have seen that there is another super robot. "LeelaPieceOdds" I wonder... is it possible to play against that robot? The robot represents different kinds of Odds. How wonderful!! Thanks to the entire Leela team for allowing us the pleasure of playing against robots.🐴
I posted about this when it first came out a month or so ago. Sure, anyone with a LiChess account can play against "LeelaPieceOdds" at any of the available odds. Note that this bot runs on a computer in Europe that is not available during the night in Europe, and also that it uses the queen odds net, so it is not quite as good at playing the various odds as the queen odds bot is at playing queen odds. But it isn't a lot worse, roughly a pawn or so worse I would say. It is mostly useful for two groups of players: 1. Strong players who can generally beat Leela Queen Odds in Rapid but who have hardly any chance with just Rook odds. I am in that group, and find that Two Knights handicap is the most appropriate for me; I can generally win in Rapid but rarely in blitz, even slow blitz. Of course we already have a Queen for Knight bot that is also suitable, but it is still quite challenging even at my level; the two knights bot is easier to defeat. 2. Players below the level needed to beat LeelaQueenOdds at the time limit they like. They can choose Queen and knight, or four minors, or Queen and Rook, or Queen and two knights or bishops, or even Queen and Two rooks! Of course, for players below 1000 or so even that won't be enough to win at Rapid, and of course as the time limit gets shorter the needed rating to win goes higher. At one minute bullet chess it might take a bullet rating over 2000 to win at queen and two rooks, there isn't much data yet. I probably couldn't do it as my own level at pure one minute bullet is far below 2000, I can't play more than about 30 moves in a minute without increment.
Thank you very much, Mr. Larry Kaufman, for this important and interesting information. Thank you for teaching us about bots. In my country I have had the opportunity to talk to some chess players and express my best concept and opinion about robots and especially LeelaOdds. There are still many players who are not aware of the existence of these fantastic rivals and sparring partners, robots that from my point of view, are a true new wonder of chess. The truth is that people have been amazed and surprised, I give the best and sincere publicity to robots and the work behind these advances that escape my knowledge, but not my understanding. Thanks again. Note. I would like to check how my new elo will be affected by LeelaQueenOdds. I have 80 new games, and my draw percentage was 42.5% playing as a white soldier. Thanks in advance. Today I am at 2290, before at 2432.
I am thinking chess is in a coin.Human beings for ever playing in one face.Now I am playing in the other face:"Antichess". Computers are as a fortres where owner forgot to close a little door behind. You must enter across this door.Forget the front.
Father
Posts: 1722
Joined: Sun Mar 19, 2006 4:39 am
Location: Colombia
Full name: Pablo Ignacio Restrepo

Re: Amazing results of LeelaKnightOdds-Dev

Post by Father »

Uri Blass wrote: Fri Jan 17, 2025 2:23 pm
Father wrote: Thu Jan 16, 2025 5:24 am
Father wrote: Wed Jan 15, 2025 11:25 pm
lkaufman wrote: Wed Jan 15, 2025 5:15 pm
Father wrote: Wed Jan 15, 2025 2:11 pm
... apparently the robots are asleep or resting, my friend LeelaQueenOdds doesn't call me, she doesn't answer my messages, it seems like she's stopped wanting to dance with me. I'm on the bench alone and despised.
The problem has been fixed. The bots were transferred to a different computer, which was missing one needed file. Should be identical to yesterday's bots now.
It could happen how things go that the prize for humans facing
LeelaQueenOdds is to occupy a place in the table of the 100 best humans. It turns out that the table of the 100 best is equivalent to a big three bagon, where there are only 100 berths for 100 people, but the number of people who want to travel on the train is substantially higher than the quota. Long live LeelaQueenOdds.!!
I wish to remember one particular game between man and machine LeelaQueenOdds. I was just watching this game, which, in my own understanding of chess, shows me the greatness that is a good draw.

The human Hissha played fearlessly, and not even the rush of time affected tactical moves in the final phase of the game, like a trapeze artist with nerves of steel. For its part, the machine made use of extraordinary resources. Such moves are worth much more than many victories. How great is the human being who invented chess, machines and robots. Life is beautiful, life is life.

[pgn][Event "Casual blitz game"]
[Site "https://lichess.org/xOR1l3OS"]
[Date "2025.01.16"]
[White "LeelaQueenOdds"]
[Black "Hissha"]
[Result "1/2-1/2"]
[UTCDate "2025.01.16"]
[UTCTime "04:02:57"]
[WhiteElo "2000"]
[BlackElo "2374"]
[WhiteTitle "BOT"]
[Variant "From Position"]
[TimeControl "60+10"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNB1KBNR w KQkq - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. g3 d5 2. Nf3 Bf5 3. d3 Nf6 4. Nc3 e6 5. h3 d4 6. Nb1 Nc6 7. g4 Bg6 8. a3 Bd6 9. Nbd2 e5 10. g5 Nd5 11. h4 Bh5 12. Ne4 Bxf3 13. exf3 Qe7 14. Bh3 Nf4 15. Bxf4 exf4 16. O-O-O Ne5 17. Rde1 O-O 18. Bf5 Nxf3 19. Ref1 Ne5 20. Rfg1 Bxa3 21. h5 Bb4 22. Rg2 a5 23. Kb1 a4 24. Ka2 a3 25. b3 b5 26. f3 c5 27. Nf6+ Kh8 28. Nxh7 c4 29. g6 cxb3+ 30. cxb3 Rfc8 31. Be4 Ra7 32. Ng5 Rac7 33. Rhh2 Rc1 34. h6 fxg6 35. hxg7+ Kxg7 36. Bd5 Rh8 37. Rxh8 Kxh8 38. Rh2+ Kg7 39. Rh7+ Kf8 40. Rh8+ Kg7 41. Rh7+ Kf8 42. Rh8+ Kg7 43. Rh7+ { The game is a draw. } 1/2-1/2[/pgn]

Looking at this game I guess the human blundered because of time trouble.
41...Kf6 was winning and maybe black was afraid from 42.Ne4+ Kf5 but black is winning after Rxe7 Bxe7(it is easy to miss the fact that the rook is defended by the bishop when you do not have a lot of time and I guess Larry missed it and did not play Kf6 because of it but I am not sure if I am right and maybe it is not the reason for Kf8).

Note that I also did a similiar mistake of missing a capture in my calculation from a winning position and lost partly because of that.
[d]r3r1k1/pp3ppp/2pb4/3p3q/6nP/1PNPP1P1/PBR2P2/3Q1RK1 b - h3 0 18
Here my calculation in the game started.

I caluclated g5 Kg2 gxh4 Rh1 h3+ and thought wrongly that I win if Rxh3 Nxe3+ fxe3 Qxd1 so I thought white needs to reply Kg1 and did not pay much attention to Rxh3(I was not 100% sure if I win easily after Kg1 so played h3+ too fast to save time for me for later moves because I assumed I may need some time after h3+ Kg1.

Only after Rxh3 I understood that my opponent will not do a stupid mistake of losing the queen so I played Qg6 that is not losing but I lost later.

Here the position that I blundered because Qg6 at that point can win the game(if Rxh4 then Nxe3+)
[d]r3r1k1/pp3p1p/2pb4/3p3q/6np/1PNPP1P1/PBR2PK1/3Q3R b - - 1 20

If I try to compare the mistakes then
Note that Larry needed 5 plies to calculate Kf6 Ne4+ Kf5 Qxe7 Bxe7 when I needed to calculate in advance 6 plies to calculate in advance
h3+ Rxh3 Nxe3+ fxe3 Qxd1 Nxd1

I consider calculating 6 plies forward in advance in 90+30 to be easier than calculating 5 plies forward in advance in 1+10 but of course larry is a better player than me and I do not know if I am right by guessing the reason for his mistake.
Uri, both your games and Larry Kaufman's that we have observed here, do not have any glaring errors... for the time controls, in my knowledge and understanding, which is no more than that of a critic and interpreter of the scientific game, they are excellent human productions of chess art... Furthermore, if we place the games of the world championships in classical time, under the magnifying glass of supercomputers, the rating will undoubtedly deteriorate for the human under a superhuman magnifying glass or beyond the human.
I am thinking chess is in a coin.Human beings for ever playing in one face.Now I am playing in the other face:"Antichess". Computers are as a fortres where owner forgot to close a little door behind. You must enter across this door.Forget the front.
Uri Blass
Posts: 10704
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Amazing results of LeelaKnightOdds-Dev

Post by Uri Blass »

Father wrote: Sat Jan 18, 2025 12:17 am
Uri, both your games and Larry Kaufman's that we have observed here, do not have any glaring errors... for the time controls, in my knowledge and understanding, which is no more than that of a critic and interpreter of the scientific game, they are excellent human productions of chess art... Furthermore, if we place the games of the world championships in classical time, under the magnifying glass of supercomputers, the rating will undoubtedly deteriorate for the human under a superhuman magnifying glass or beyond the human.

I disagree.
The games have tactical calculation mistakes.

In the case of my loss against human opponent I did not show the full game but only part of it when I made the mistake.
In my case I can explain pshychological reason for playing the blunder h3+ too fast because my opponent did not play a move that I considered illogical so I did not suspect that I miss something and also I had no alternative that seems to me good after not thinking much(I thought that without my h pawn white has the advantage because of better pawn structure so I did not consider except h3+ other move except Be7 that I rejected as too passive).

[d]r3r1k1/pp3p1p/2pb4/3p3q/6np/1PNPP1P1/PBR2PK1/3Q3R b - - 1 20

The unlucky thing for me is the fact that white can reply Kg1 against h3+ so the main time that I used was to look at this line earlier and did not check the line h3+ Rxh3 to see if it is correct.

In case white had no move against h3+ except Rxh3 I probably could see that Rxh3 is strong because I could check Rxh3 for more time and see that the line is bad for me and think more about other alternatives and maybe could find Qg6(cannot check it)

My opponent told me after the game that he saw the full lines with Nxd1 some moves before I played h3+ and also saw the line that I did not play Qg6 Rxh4 Nxf2(Engines claim Nxe3+ is better but Nxf2 is also winning) but hoped he is going to find something else after Qg6.

He played the last moves relatively fast because he understood he had no good alternatives and it was also a good practical decision because I had less time to think on his time.