We are reaching the Max Odds that a top engine can give the top 3 GM's

[Spliffjiffer]

...through down its pants, chris !!!

Yes Chris - straight down the dragon's pants!

It's what the watching Europeans expect of you!

ahm, yes mayby thats right but a bit overestimated...,in gernan(y) we do not better than you in britain, or england, i think we can say that this is not an inferiority for us even if i do not speak in your own language...if u wanna tell im not on a level with U thats probably correct because im even not on at same level with my german friends...so its ok....i think if we do not speak the same language we are no foes...thats mportant imo...i i like to have you brexits

[lkaufman]

I have a range of PCs. From 64x threadripper to 6x i7 laptop through some 6x i7 PCs, all with GPU, but I don’t suppose Dragon is using GPU?
In practice, I quite like the iPhone, because then I can play a few moves on the cafe or whatever. I’m pretty unlikely to sit down for a serious time controlled session - I get too many interruptions.
So, probably the laptop. I guess, being reasonable, can human hold against computer, I’ld suggest “friendly” correspondence conditions for me and game in 15 for you. I’m pretty busy right now.
Alternatively, play move by move in public forum. I’ll have to trust you to stick to move as game in 15, and you’ll have to trust me to not ask Stockfish. Anyway, you could tell if I was using comp help. And kibitzers could keep their mouths shut.
Ed’s forum? I’m sure he could set something up.

One way to do this would be on chess.com. You'd have to have an account, but a free one should work. You would be playing against Dragon on my 32 core Threadripper with book, at knight odds. You can pick any long time limit you like, enough to accommodate any likely interruptions, and I can set Dragon internally to play as if it were just a one minute game plus ten second increment. I do trust you not to use computer help. We would just have to agree on when to play the game or games. If you prefer a move by move game, we can talk about that.

You built a knight odds book? Surely the idea is to discover whether a strong chess player thinking on his own without time pressure is better than/same as a computer thinking on its own. With a special book, I guess selecting lines that it has generally done better with over gazillions of test games, doesn’t seem quite right to me. It’s not as if humans have odds books all prepared. Why an engine should have one?
We would want to see raw talent and raw strength right from the start. Do you have a problem operating without a specially prepared book?
Move by move works fine on a forum thread. It also allows a degree of oversight by onlookers. I don’t mind to allow you a 32 core threadripper if you think you need it.

The knight odds opening book is needed for matches against grandmasters, both to provide sufficient opening variety and because they are allowed to prepare for the match by analyzing the opening with engines. But if you are not doing that, then we have no reason to use the book. It's not a book tuned by engine games, rather it's just based on its own analysis with modifications based on games against myself and other titled players. The main goal is to avoid human preparation, but no need in your case. If we play move by move, I would just use my 10 core i9 laptop for the game, since it's not practical for us to use the Threadripper in this way, it's needed for development. I can set it for a fixed time per move, maybe 20 seconds, rather than game/15 as I might have to restart and would lose the time info.

OK

We can play here on talkchess if you like. I rather expect you to win if you spend much time on the game, it all depends on what you mean by "friendly" correspondence. Clearly we can't give knight odds to a strong player who is taking 20 minutes per move or so, but I don't think that's your intention. Just let us know how much time you actually take (roughly) so we can take that into consideration when evaluating the game or games afterwards. I hope we can trust the readers of this forum not to post advice, especially advice based on computer analysis. We clearly can't give Stockfish knight odds!

Seems reasonable. I guess it will be casually serious, so to speak. Instead of going to the cafe and posting social media, I’ll open the chess board (on a forum will be fine, probably we can use here plus Ed’s, mirroring. I didn’t forget I got banned from posting here once for no good reason while playing one of the mods (Robin Smith), thus stopping the game.
I’m very unlikely to spend more than one minute a move, frankly, mostly less. As you no doubt realise I’ll just play positional with trying to avoid tactics and tactical positions. Moves will be intuitive with tactics over-checking, so it doesn’t really make sense to think about “time” in this positional human sense. Strong players don’t gain 70 Elo by doubling their time allowance etc etc. I’m mostly interested in demonstrating (maybe I’ll be wrong of course) that any strong player who a) hadn’t got a clock stressing him and b) has space to think a bit and c) doesn’t blunder (which is always possible too) is going to wipe out any engine with knight odds. As you more or less already concede.

What I suggest is that I play the black side of the two Alekhine possibilities. Then we can see if your engine without external book, presses, likes space, prefers to hang back behind a smaller centre or whatever.

First off:
e4, remove nb1, nf6, bookless white to move .....

rnbqkb1r/pppppppp/5n2/4P3/8/8/PPPP1PPP/R1BQKBNR b KQkq - 0 1

Okay, fair enough. Dragon 2 MCTS (with my settings) replies 2.e5 as expected after 20 seconds. We can make this a separate post or keep it here, as you like. I don't remember the precise procedure to make it display the board here properly, maybe someone can remind me. No book, but it wouldn't have mattered anyway, as the Alekhine is not even in the knight odds book. Actually, Dragon (with or without book) wouldn't open 1.e4 with the b1 knight missing, as the reply 1...d5! is just too strong (known since the 1800s), but since you committed to 1.e4 Nf6 there is no objection to 1.e4. I don't agree with your claim that strong players don't gain something like 70 Elo by doubling their time (at least up to 4 min per move or so) or that blundering is a yes or no thing; at 30" or so per move top GMs make many small blunders (typically ones that will lose a pawn or suffer some significant positional cost) which they would rarely do at 3' per move. I know that the quality of my standard tournament games is vastly superior to the quality of my Rapid games, and I think this is generally true of GMs and IMs, but not for all strong players, some are just more intuitive and don't rely so much on calculation. Based on your comments, I think you belong to this second group. But we don't need to agree on this point, let's just enjoy the game!

[chrisw]

I have a range of PCs. From 64x threadripper to 6x i7 laptop through some 6x i7 PCs, all with GPU, but I don’t suppose Dragon is using GPU?
In practice, I quite like the iPhone, because then I can play a few moves on the cafe or whatever. I’m pretty unlikely to sit down for a serious time controlled session - I get too many interruptions.
So, probably the laptop. I guess, being reasonable, can human hold against computer, I’ld suggest “friendly” correspondence conditions for me and game in 15 for you. I’m pretty busy right now.
Alternatively, play move by move in public forum. I’ll have to trust you to stick to move as game in 15, and you’ll have to trust me to not ask Stockfish. Anyway, you could tell if I was using comp help. And kibitzers could keep their mouths shut.
Ed’s forum? I’m sure he could set something up.

One way to do this would be on chess.com. You'd have to have an account, but a free one should work. You would be playing against Dragon on my 32 core Threadripper with book, at knight odds. You can pick any long time limit you like, enough to accommodate any likely interruptions, and I can set Dragon internally to play as if it were just a one minute game plus ten second increment. I do trust you not to use computer help. We would just have to agree on when to play the game or games. If you prefer a move by move game, we can talk about that.

You built a knight odds book? Surely the idea is to discover whether a strong chess player thinking on his own without time pressure is better than/same as a computer thinking on its own. With a special book, I guess selecting lines that it has generally done better with over gazillions of test games, doesn’t seem quite right to me. It’s not as if humans have odds books all prepared. Why an engine should have one?
We would want to see raw talent and raw strength right from the start. Do you have a problem operating without a specially prepared book?
Move by move works fine on a forum thread. It also allows a degree of oversight by onlookers. I don’t mind to allow you a 32 core threadripper if you think you need it.

The knight odds opening book is needed for matches against grandmasters, both to provide sufficient opening variety and because they are allowed to prepare for the match by analyzing the opening with engines. But if you are not doing that, then we have no reason to use the book. It's not a book tuned by engine games, rather it's just based on its own analysis with modifications based on games against myself and other titled players. The main goal is to avoid human preparation, but no need in your case. If we play move by move, I would just use my 10 core i9 laptop for the game, since it's not practical for us to use the Threadripper in this way, it's needed for development. I can set it for a fixed time per move, maybe 20 seconds, rather than game/15 as I might have to restart and would lose the time info.

OK

We can play here on talkchess if you like. I rather expect you to win if you spend much time on the game, it all depends on what you mean by "friendly" correspondence. Clearly we can't give knight odds to a strong player who is taking 20 minutes per move or so, but I don't think that's your intention. Just let us know how much time you actually take (roughly) so we can take that into consideration when evaluating the game or games afterwards. I hope we can trust the readers of this forum not to post advice, especially advice based on computer analysis. We clearly can't give Stockfish knight odds!

Seems reasonable. I guess it will be casually serious, so to speak. Instead of going to the cafe and posting social media, I’ll open the chess board (on a forum will be fine, probably we can use here plus Ed’s, mirroring. I didn’t forget I got banned from posting here once for no good reason while playing one of the mods (Robin Smith), thus stopping the game.
I’m very unlikely to spend more than one minute a move, frankly, mostly less. As you no doubt realise I’ll just play positional with trying to avoid tactics and tactical positions. Moves will be intuitive with tactics over-checking, so it doesn’t really make sense to think about “time” in this positional human sense. Strong players don’t gain 70 Elo by doubling their time allowance etc etc. I’m mostly interested in demonstrating (maybe I’ll be wrong of course) that any strong player who a) hadn’t got a clock stressing him and b) has space to think a bit and c) doesn’t blunder (which is always possible too) is going to wipe out any engine with knight odds. As you more or less already concede.

What I suggest is that I play the black side of the two Alekhine possibilities. Then we can see if your engine without external book, presses, likes space, prefers to hang back behind a smaller centre or whatever.

First off:
e4, remove nb1, nf6, bookless white to move .....

rnbqkb1r/pppppppp/5n2/4P3/8/8/PPPP1PPP/R1BQKBNR b KQkq - 0 1

Okay, fair enough. Dragon 2 MCTS (with my settings) replies 2.e5 as expected after 20 seconds. We can make this a separate post or keep it here, as you like. I don't remember the precise procedure to make it display the board here properly, maybe someone can remind me. No book, but it wouldn't have mattered anyway, as the Alekhine is not even in the knight odds book. Actually, Dragon (with or without book) wouldn't open 1.e4 with the b1 knight missing, as the reply 1...d5! is just too strong (known since the 1800s), but since you committed to 1.e4 Nf6 there is no objection to 1.e4. I don't agree with your claim that strong players don't gain something like 70 Elo by doubling their time (at least up to 4 min per move or so) or that blundering is a yes or no thing; at 30" or so per move top GMs make many small blunders (typically ones that will lose a pawn or suffer some significant positional cost) which they would rarely do at 3' per move. I know that the quality of my standard tournament games is vastly superior to the quality of my Rapid games, and I think this is generally true of GMs and IMs, but not for all strong players, some are just more intuitive and don't rely so much on calculation. Based on your comments, I think you belong to this second group. But we don't need to agree on this point, let's just enjoy the game!

With dyslexia and no board to look at, I’ll posit Nd5

I assume you have pondering turned off?

[lkaufman]

Seems reasonable. I guess it will be casually serious, so to speak. Instead of going to the cafe and posting social media, I’ll open the chess board (on a forum will be fine, probably we can use here plus Ed’s, mirroring. I didn’t forget I got banned from posting here once for no good reason while playing one of the mods (Robin Smith), thus stopping the game.
I’m very unlikely to spend more than one minute a move, frankly, mostly less. As you no doubt realise I’ll just play positional with trying to avoid tactics and tactical positions. Moves will be intuitive with tactics over-checking, so it doesn’t really make sense to think about “time” in this positional human sense. Strong players don’t gain 70 Elo by doubling their time allowance etc etc. I’m mostly interested in demonstrating (maybe I’ll be wrong of course) that any strong player who a) hadn’t got a clock stressing him and b) has space to think a bit and c) doesn’t blunder (which is always possible too) is going to wipe out any engine with knight odds. As you more or less already concede.

What I suggest is that I play the black side of the two Alekhine possibilities. Then we can see if your engine without external book, presses, likes space, prefers to hang back behind a smaller centre or whatever.

First off:
e4, remove nb1, nf6, bookless white to move .....

rnbqkb1r/pppppppp/5n2/4P3/8/8/PPPP1PPP/R1BQKBNR b KQkq - 0 1

Okay, fair enough. Dragon 2 MCTS (with my settings) replies 2.e5 as expected after 20 seconds. We can make this a separate post or keep it here, as you like. I don't remember the precise procedure to make it display the board here properly, maybe someone can remind me. No book, but it wouldn't have mattered anyway, as the Alekhine is not even in the knight odds book. Actually, Dragon (with or without book) wouldn't open 1.e4 with the b1 knight missing, as the reply 1...d5! is just too strong (known since the 1800s), but since you committed to 1.e4 Nf6 there is no objection to 1.e4. I don't agree with your claim that strong players don't gain something like 70 Elo by doubling their time (at least up to 4 min per move or so) or that blundering is a yes or no thing; at 30" or so per move top GMs make many small blunders (typically ones that will lose a pawn or suffer some significant positional cost) which they would rarely do at 3' per move. I know that the quality of my standard tournament games is vastly superior to the quality of my Rapid games, and I think this is generally true of GMs and IMs, but not for all strong players, some are just more intuitive and don't rely so much on calculation. Based on your comments, I think you belong to this second group. But we don't need to agree on this point, let's just enjoy the game!
[/quote]

With dyslexia and no board to look at, I’ll posit Nd5

I assume you have pondering turned off?
[/quote]

After 20" Dragon chose the normal looking 3.d4 after 2...Nd5. If I figure out how to post the position properly I'll do so. I exit the GUI and use the computer for other things between moves, no pondering.

[lkaufman]

[pgn] [Event "Knight Odds Match"]
[Site "talkchess"]
[Date "2021.05.20"]
[Round "1"]
[White "KomodoDragon2 MCTS"]
[Black "Whittington, Chris"]
[Result "*"]
[Annotator "Kaufman,Larry"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/R1BQKBNR w KQkq - 0 1"]
[PlyCount "5"]
[EventDate "2021.05.20"]

{[#]} 1. e4 Nf6 2. e5 Nd5 3. d4 *[/pgn]

[chrisw]

[pgn] [Event "Knight Odds Match"]
[Site "talkchess"]
[Date "2021.05.20"]
[Round "1"]
[White "KomodoDragon2 MCTS"]
[Black "Whittington, Chris"]
[Result "*"]
[Annotator "Kaufman,Larry"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/R1BQKBNR w KQkq - 0 1"]
[PlyCount "5"]
[EventDate "2021.05.20"]

{[#]} 1. e4 Nf6 2. e5 Nd5 3. d4 d6 [/pgn]

d6

[Chessqueen]

[pgn] [Event "Knight Odds Match"]
[Site "talkchess"]
[Date "2021.05.20"]
[Round "1"]
[White "KomodoDragon2 MCTS"]
[Black "Whittington, Chris"]
[Result "*"]
[Annotator "Kaufman,Larry"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/R1BQKBNR w KQkq - 0 1"]
[PlyCount "5"]
[EventDate "2021.05.20"]

{[#]} 1. e4 Nf6 2. e5 Nd5 3. d4 d6 [/pgn]

d6

Both of you can use this Windows 10 App, by simply click on start NOW 2nd Download and login
https://download.cnet.com/Real-Chess-On ... 72866.html

[lkaufman]

[pgn] [Event "Knight Odds Match"]
[Site "talkchess"]
[Date "2021.05.20"]
[Round "1"]
[White "KomodoDragon2 MCTS"]
[Black "Whittington, Chris"]
[Result "*"]
[Annotator "Kaufman,Larry"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/R1BQKBNR w KQkq - 0 1"]
[PlyCount "5"]
[EventDate "2021.05.20"]

{[#]} 1. e4 Nf6 2. e5 Nd5 3. d4 d6 [/pgn]

d6

[pgn] [Event "?"]
[Site "?"]
[Date "2021.05.16"]
[Round "1"]
[White "KomodoDragon2 MCTS"]
[Black "Whittington, Chris"]
[Result "*"]
[Annotator "Kaufman,Larry"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/R1BQKBNR w KQkq - 0 1"]
[PlyCount "7"]
[EventDate "2021.??.??"]

{[#]} 1. e4 Nf6 2. e5 Nd5 3. d4 d6 4. Nf3 *

[/pgn]

[Chessqueen]

[pgn] [Event "Knight Odds Match"]
[Site "talkchess"]
[Date "2021.05.20"]
[Round "1"]
[White "KomodoDragon2 MCTS"]
[Black "Whittington, Chris"]
[Result "*"]
[Annotator "Kaufman,Larry"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/R1BQKBNR w KQkq - 0 1"]
[PlyCount "5"]
[EventDate "2021.05.20"]

{[#]} 1. e4 Nf6 2. e5 Nd5 3. d4 d6 [/pgn]

d6

[pgn] [Event "?"]
[Site "?"]
[Date "2021.05.16"]
[Round "1"]
[White "KomodoDragon2 MCTS"]
[Black "Whittington, Chris"]
[Result "*"]
[Annotator "Kaufman,Larry"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/R1BQKBNR w KQkq - 0 1"]
[PlyCount "7"]
[EventDate "2021.??.??"]

{[#]} 1. e4 Nf6 2. e5 Nd5 3. d4 d6 4. Nf3 *

[/pgn]

At that pace this game will finish by the end of June 2021

[chrisw]

[pgn] [Event "Knight Odds Match"]
[Site "talkchess"]
[Date "2021.05.20"]
[Round "1"]
[White "KomodoDragon2 MCTS"]
[Black "Whittington, Chris"]
[Result "*"]
[Annotator "Kaufman,Larry"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/R1BQKBNR w KQkq - 0 1"]
[PlyCount "5"]
[EventDate "2021.05.20"]

{[#]} 1. e4 Nf6 2. e5 Nd5 3. d4 d6 [/pgn]

d6

[pgn] [Event "?"]
[Site "?"]
[Date "2021.05.16"]
[Round "1"]
[White "KomodoDragon2 MCTS"]
[Black "Whittington, Chris"]
[Result "*"]
[Annotator "Kaufman,Larry"]
[SetUp "1"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/R1BQKBNR w KQkq - 0 1"]
[PlyCount "7"]
[EventDate "2021.??.??"]

{[#]} 1. e4 Nf6 2. e5 Nd5 3. d4 d6 4. Nf3 Bg4

[/pgn]

Bg4