I toast to LeelaqueenOdds and its creators. I raise my wine glass high and shout: Long live the queen. We have a queen!!

[lkaufman]

Today IM Levy Rozman, the most popular chess streamer, played a five game match with LeelaQueenOdds at 5'3", winning by 3 to 2. Yesterday and today LQO is playing about 7000 games per day, maybe comparable to all the other bots on LiChess combined?! Just incredible interest. Meanwhile the knight odds bot defeated another top player, Alexei Sarana, by 12.5 to 0.5 in blitz (3'1"). Levy plans to play rook odds next.

A comet in the sky.
"It's a bird!!"
"It's a plane!!"
"Nooooo!!! It's IM wateenellende"
Superhuman sweeps LeelaQueenOdds 1-0!!

After a hundred games, he has dropped below 40%, but that is still far better than anyone else, I think the previous record was about 25%. But strangely, a Russian GM, Andrey Belozerov, FIDE 2400, playing the same bot at a leisurely 3'2" TC, lost by an astounding 14.5 to 0.5! I can't make any sense of that, it is absurd.

Good evening, Mr. Larry Kaufman. Expertise playing against chess machines seems to be one of the determining factors in obtaining certain desired results. Without a doubt, if you add being a Master to this, there will be more unbalancing elements. LeelaQueenOdds has dealt some beatings to advanced chess players with a bullet rating of 2303, where the machine beats the human 61.5 to 0.5 in CR 1+0. A few years ago, it seemed like pure and simple "science fiction." But no! It's reality in today's world.

Well, Leela can flag players at 1'0" so a shutout isn't so hard to imagine, but at 3'2" to win 14 games plus a draw out of 15 vs a GM still rated 2400 FIDE seems absolutely impossible. Anyway, today Hikaru joined the ranks of players taking on the Leela Queen Odds one minute challenge. He split 4 games alternating colors (so already not a fair comparison to the many top players who follow tradition and take Black), then gave himself White for a "tiebreaker" and won and called that a match win. Probably he didn't know that Leela was only trained to play White so only those games are relevant, so at 1 to 1 we don't know anything. He says he plans to play knight odds next, hopefully at a long enough time control to give him some chance.

[Father]

Today IM Levy Rozman, the most popular chess streamer, played a five game match with LeelaQueenOdds at 5'3", winning by 3 to 2. Yesterday and today LQO is playing about 7000 games per day, maybe comparable to all the other bots on LiChess combined?! Just incredible interest. Meanwhile the knight odds bot defeated another top player, Alexei Sarana, by 12.5 to 0.5 in blitz (3'1"). Levy plans to play rook odds next.

A comet in the sky.
"It's a bird!!"
"It's a plane!!"
"Nooooo!!! It's IM wateenellende"
Superhuman sweeps LeelaQueenOdds 1-0!!

After a hundred games, he has dropped below 40%, but that is still far better than anyone else, I think the previous record was about 25%. But strangely, a Russian GM, Andrey Belozerov, FIDE 2400, playing the same bot at a leisurely 3'2" TC, lost by an astounding 14.5 to 0.5! I can't make any sense of that, it is absurd.

Good evening, Mr. Larry Kaufman. Expertise playing against chess machines seems to be one of the determining factors in obtaining certain desired results. Without a doubt, if you add being a Master to this, there will be more unbalancing elements. LeelaQueenOdds has dealt some beatings to advanced chess players with a bullet rating of 2303, where the machine beats the human 61.5 to 0.5 in CR 1+0. A few years ago, it seemed like pure and simple "science fiction." But no! It's reality in today's world.

Well, Leela can flag players at 1'0" so a shutout isn't so hard to imagine, but at 3'2" to win 14 games plus a draw out of 15 vs a GM still rated 2400 FIDE seems absolutely impossible. Anyway, today Hikaru joined the ranks of players taking on the Leela Queen Odds one minute challenge. He split 4 games alternating colors (so already not a fair comparison to the many top players who follow tradition and take Black), then gave himself White for a "tiebreaker" and won and called that a match win. Probably he didn't know that Leela was only trained to play White so only those games are relevant, so at 1 to 1 we don't know anything. He says he plans to play knight odds next, hopefully at a long enough time control to give him some chance.

Mr. Larry Kaufman, did Nakamura play LeelaQueenOdds? Can I watch the matches? It's starting to become a reality, in the miracle I believe will happen: to see the Top 20 in combat against the queen of cyber chess.

[Father]

Today IM Levy Rozman, the most popular chess streamer, played a five game match with LeelaQueenOdds at 5'3", winning by 3 to 2. Yesterday and today LQO is playing about 7000 games per day, maybe comparable to all the other bots on LiChess combined?! Just incredible interest. Meanwhile the knight odds bot defeated another top player, Alexei Sarana, by 12.5 to 0.5 in blitz (3'1"). Levy plans to play rook odds next.

A comet in the sky.
"It's a bird!!"
"It's a plane!!"
"Nooooo!!! It's IM wateenellende"
Superhuman sweeps LeelaQueenOdds 1-0!!

After a hundred games, he has dropped below 40%, but that is still far better than anyone else, I think the previous record was about 25%. But strangely, a Russian GM, Andrey Belozerov, FIDE 2400, playing the same bot at a leisurely 3'2" TC, lost by an astounding 14.5 to 0.5! I can't make any sense of that, it is absurd.

Good evening, Mr. Larry Kaufman. Expertise playing against chess machines seems to be one of the determining factors in obtaining certain desired results. Without a doubt, if you add being a Master to this, there will be more unbalancing elements. LeelaQueenOdds has dealt some beatings to advanced chess players with a bullet rating of 2303, where the machine beats the human 61.5 to 0.5 in CR 1+0. A few years ago, it seemed like pure and simple "science fiction." But no! It's reality in today's world.

Well, Leela can flag players at 1'0" so a shutout isn't so hard to imagine, but at 3'2" to win 14 games plus a draw out of 15 vs a GM still rated 2400 FIDE seems absolutely impossible. Anyway, today Hikaru joined the ranks of players taking on the Leela Queen Odds one minute challenge. He split 4 games alternating colors (so already not a fair comparison to the many top players who follow tradition and take Black), then gave himself White for a "tiebreaker" and won and called that a match win. Probably he didn't know that Leela was only trained to play White so only those games are relevant, so at 1 to 1 we don't know anything. He says he plans to play knight odds next, hopefully at a long enough time control to give him some chance.

Mr. Larry Kaufman, I am personally convinced that if GMs and Masters in general play consistently against LeelaQueenOdds, they will and will continue to enhance their epic qualities of being Masters. But it is also a "human gesture" that "exalts the greatness of possessing the quality of being a person" and "makes us happy" as "the mass of fans." I sincerely thank all the Masters and, from "my little wooden horse," invite them to ride together to achieve their dreams and ideals in the cavalcade of life, without distinction of any nature, all together. This time I raise the imaginary glass over my little horse Catecan and toast to chess to which my heart remains tied, the heart of my horse and the life-giving thread with which Our "giver of life" charges our souls with energy. I hope with faith, "the great Mozart of chess: Magnum Carlsen" ... perhaps before I leave with the friend of my little wooden horse: the great Pegasus on my last flight to the deep universe.

[lkaufman]

Today IM Levy Rozman, the most popular chess streamer, played a five game match with LeelaQueenOdds at 5'3", winning by 3 to 2. Yesterday and today LQO is playing about 7000 games per day, maybe comparable to all the other bots on LiChess combined?! Just incredible interest. Meanwhile the knight odds bot defeated another top player, Alexei Sarana, by 12.5 to 0.5 in blitz (3'1"). Levy plans to play rook odds next.

A comet in the sky.
"It's a bird!!"
"It's a plane!!"
"Nooooo!!! It's IM wateenellende"
Superhuman sweeps LeelaQueenOdds 1-0!!

After a hundred games, he has dropped below 40%, but that is still far better than anyone else, I think the previous record was about 25%. But strangely, a Russian GM, Andrey Belozerov, FIDE 2400, playing the same bot at a leisurely 3'2" TC, lost by an astounding 14.5 to 0.5! I can't make any sense of that, it is absurd.

Good evening, Mr. Larry Kaufman. Expertise playing against chess machines seems to be one of the determining factors in obtaining certain desired results. Without a doubt, if you add being a Master to this, there will be more unbalancing elements. LeelaQueenOdds has dealt some beatings to advanced chess players with a bullet rating of 2303, where the machine beats the human 61.5 to 0.5 in CR 1+0. A few years ago, it seemed like pure and simple "science fiction." But no! It's reality in today's world.

Well, Leela can flag players at 1'0" so a shutout isn't so hard to imagine, but at 3'2" to win 14 games plus a draw out of 15 vs a GM still rated 2400 FIDE seems absolutely impossible. Anyway, today Hikaru joined the ranks of players taking on the Leela Queen Odds one minute challenge. He split 4 games alternating colors (so already not a fair comparison to the many top players who follow tradition and take Black), then gave himself White for a "tiebreaker" and won and called that a match win. Probably he didn't know that Leela was only trained to play White so only those games are relevant, so at 1 to 1 we don't know anything. He says he plans to play knight odds next, hopefully at a long enough time control to give him some chance.

Mr. Larry Kaufman, did Nakamura play LeelaQueenOdds? Can I watch the matches? It's starting to become a reality, in the miracle I believe will happen: to see the Top 20 in combat against the queen of cyber chess.

Search for games against "GamingWarlord", or just watch his video.

[lkaufman]

Mr. Larry Kaufman, I am personally convinced that if GMs and Masters in general play consistently against LeelaQueenOdds, they will and will continue to enhance their epic qualities of being Masters. But it is also a "human gesture" that "exalts the greatness of possessing the quality of being a person" and "makes us happy" as "the mass of fans." I sincerely thank all the Masters and, from "my little wooden horse," invite them to ride together to achieve their dreams and ideals in the cavalcade of life, without distinction of any nature, all together. This time I raise the imaginary glass over my little horse Catecan and toast to chess to which my heart remains tied, the heart of my horse and the life-giving thread with which Our "giver of life" charges our souls with energy. I hope with faith, "the great Mozart of chess: Magnum Carlsen" ... perhaps before I leave with the friend of my little wooden horse: the great Pegasus on my last flight to the deep universe.

Carlsen is widely and properly considered the greatest player of all in classical chess, Rapid chess, and even blitz chess, but Nakamura is generally considered to be even stronger in bullet chess, so you have already gotten the best possible opponent for LQO in bullet.

[Father]

Mr. Larry Kaufman, I am personally convinced that if GMs and Masters in general play consistently against LeelaQueenOdds, they will and will continue to enhance their epic qualities of being Masters. But it is also a "human gesture" that "exalts the greatness of possessing the quality of being a person" and "makes us happy" as "the mass of fans." I sincerely thank all the Masters and, from "my little wooden horse," invite them to ride together to achieve their dreams and ideals in the cavalcade of life, without distinction of any nature, all together. This time I raise the imaginary glass over my little horse Catecan and toast to chess to which my heart remains tied, the heart of my horse and the life-giving thread with which Our "giver of life" charges our souls with energy. I hope with faith, "the great Mozart of chess: Magnum Carlsen" ... perhaps before I leave with the friend of my little wooden horse: the great Pegasus on my last flight to the deep universe.

Carlsen is widely and properly considered the greatest player of all in classical chess, Rapid chess, and even blitz chess, but Nakamura is generally considered to be even stronger in bullet chess, so you have already gotten the best possible opponent for LQO in bullet.

Thank you, Mr. Larry Kaufman. "Literally." "This is incredible." For me, "it's beyond fiction." What was the final result? What follows is "the perfecting of the robot."
Thanks again to the Leela Team for allowing us to use this robot!

[Brunetti]

Hi Larry,

a thought: unlike any other modern engine, for Leela Queen Odds we now have thousands of games (on Lichess) at various time controls against known human players. This would finally make it possible to provide a much more accurate Elo rating on the human scale for all other current engines, if only they could play matches against this program.

So, I’d like to ask if there are any plans for its release (or if it has already been released and I’m just unaware of it), perhaps even an earlier version that has still played thousands of games. If so, it would be very interesting for the purpose I mentioned.

Thanks,
Alex

[lkaufman]

Mr. Larry Kaufman, I am personally convinced that if GMs and Masters in general play consistently against LeelaQueenOdds, they will and will continue to enhance their epic qualities of being Masters. But it is also a "human gesture" that "exalts the greatness of possessing the quality of being a person" and "makes us happy" as "the mass of fans." I sincerely thank all the Masters and, from "my little wooden horse," invite them to ride together to achieve their dreams and ideals in the cavalcade of life, without distinction of any nature, all together. This time I raise the imaginary glass over my little horse Catecan and toast to chess to which my heart remains tied, the heart of my horse and the life-giving thread with which Our "giver of life" charges our souls with energy. I hope with faith, "the great Mozart of chess: Magnum Carlsen" ... perhaps before I leave with the friend of my little wooden horse: the great Pegasus on my last flight to the deep universe.

Carlsen is widely and properly considered the greatest player of all in classical chess, Rapid chess, and even blitz chess, but Nakamura is generally considered to be even stronger in bullet chess, so you have already gotten the best possible opponent for LQO in bullet.

Thank you, Mr. Larry Kaufman. "Literally." "This is incredible." For me, "it's beyond fiction." What was the final result? What follows is "the perfecting of the robot."
Thanks again to the Leela Team for allowing us to use this robot!

So far Hikaru just played five games, all 1+0 bullet. He played as Black twice, splitting 1 to 1, and played as White three times, winning 2 to 1. I hope he will play more games; so far he has only said that he will play some games with Leela Knight Odds.

[lkaufman]

Hi Larry,

a thought: unlike any other modern engine, for Leela Queen Odds we now have thousands of games (on Lichess) at various time controls against known human players. This would finally make it possible to provide a much more accurate Elo rating on the human scale for all other current engines, if only they could play matches against this program.

So, I’d like to ask if there are any plans for its release (or if it has already been released and I’m just unaware of it), perhaps even an earlier version that has still played thousands of games. If so, it would be very interesting for the purpose I mentioned.

Thanks,
Alex

The queen odds net has already been released (see Lc0 on Discord to ask how to get it), the knight/rook odds net has not been released. We do have quite accurate stats for LQO. Tonight I calculated that the performance of LQO playing White at 1'+1" against known FIDE titled (FM and up) players since the major Feb.27 upgrade is 2547 FIDE vs GMs and 2575 vs IMs and FMs. I would trust the results against the GMs more (since nearly equal results) so I think you can use 2550 as a fair estimate of the break-even FIDE rating at 1'1". But note that this net is optimized to play against "humanized" nets, so the best use of it should be to rate the various human-like nets like Maia and Elite Leela at node counts that give them about equal chances vs. the settings used in human play, then treat those as reliably rated vs humans at 1'1". You can try to do the same with longer time control data, but once you get beyond blitz you won't have enough games against titled players, basically the only ones on LiChess where you can expect to find a reliable FIDE rating.

[Father]

Hi Larry,

a thought: unlike any other modern engine, for Leela Queen Odds we now have thousands of games (on Lichess) at various time controls against known human players. This would finally make it possible to provide a much more accurate Elo rating on the human scale for all other current engines, if only they could play matches against this program.

So, I’d like to ask if there are any plans for its release (or if it has already been released and I’m just unaware of it), perhaps even an earlier version that has still played thousands of games. If so, it would be very interesting for the purpose I mentioned.

Thanks,
Alex

The queen odds net has already been released (see Lc0 on Discord to ask how to get it), the knight/rook odds net has not been released. We do have quite accurate stats for LQO. Tonight I calculated that the performance of LQO playing White at 1'+1" against known FIDE titled (FM and up) players since the major Feb.27 upgrade is 2547 FIDE vs GMs and 2575 vs IMs and FMs. I would trust the results against the GMs more (since nearly equal results) so I think you can use 2550 as a fair estimate of the break-even FIDE rating at 1'1". But note that this net is optimized to play against "humanized" nets, so the best use of it should be to rate the various human-like nets like Maia and Elite Leela at node counts that give them about equal chances vs. the settings used in human play, then treat those as reliably rated vs humans at 1'1". You can try to do the same with longer time control data, but once you get beyond blitz you won't have enough games against titled players, basically the only ones on LiChess where you can expect to find a reliable FIDE rating.

Good evening, Alex. I hope you are well and continue to be well. Life has gifted me with 64 years of age. Beginning in 1980, I had my first exposure to a chess computer: "Chess Challenger Level 7" from Fidelity Electronics. At the time, at 19 years old, it was an unimaginable dream. Then came the passion to discover "the perfection of chess itself, which flourished from computers and made up for my own imperfections." Computers filled the gaps left by the inability to face Masters, and each day I became more and more familiar with machines. What I then considered interaction with computers was, a little while before, "science fiction" for my elders. Then,
with a good friend who is equally passionate about chess and computers, we talked and wondered if one day there would be a computer that, taking just one second to respond, could beat all humans. We estimated the year 2040 for this. We calculated speed, brute force, and selective systems. At that time, it wasn't reasonable to imagine a computer playing without a queen... "It was too much!" It was a bit like asking for pears from an elm tree.
Last September, Mr. Larry Kaufman introduced me to the Odds robots, and based on my own experience, "it represents a benchmark in software, like the Apollo from Earth to the Moon." Congratulations to everyone who has worked on this through science, and congratulations to astronaut Hikaru Nakamura for his great flight.