mmt Vs. Ovyron (G4 D5 BG2)

Discussion of computer chess matches and engine tournaments.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
mmt
Posts: 210
Joined: Sun Aug 25, 2019 6:33 am
Full name: .

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by mmt » Fri Feb 28, 2020 2:31 am

If axb4 then axb4.

jp
Posts: 1254
Joined: Mon Apr 23, 2018 5:54 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by jp » Fri Feb 28, 2020 2:55 am

mmt wrote:
Fri Feb 28, 2020 12:59 am
4r1k1/2qnrpp1/2p3n1/p2p2Np/1P1P3P/P2QPP2/2N2K2/5R1R b - - 0 1
LC0 at -2.53. SF at -3.41.
Komodo 13.2, depth 27, top 3 moves: -1.31 (...axb4), -1.23, -1.17.

Uri Blass wrote:
Thu Feb 27, 2020 8:05 am
I see no way to find practically that a move is the only winning move.
Here is a simple challange.

4b1k1/8/5n1r/4N3/8/5Q2/3K4/8 w - - 0 1

Can you practically find the winning move and prove the other moves are a draw when you do not use the syzygy tablebases that have the solution.

I did not try it but I believe that it is not a simple problem.
Maybe unassisted engine without tablebases can find the right line and I did not test it.
This may be a good test of Ovyron's wild claim he can magically tag all non-winning lines.

But for obvious reasons it may not be a good test that he "needs" a hardware upgrade, unless a TB-free engine can find the right line (doubtful at the moment, but maybe if you go to depth 100 or 120?).


Zenmastur wrote:
Thu Feb 27, 2020 11:16 am
The difference between this test and a “normal” position from a game is that it's intuitively obvious when the goal has or hasn't been reached.
But unfortunately that's not the only difference, as I said in my post about hypothetical special skills for pruning.


Ovyron wrote:
Thu Feb 27, 2020 11:34 pm
I have become an expert in predicting opponent's moves, so I can predict what faster hardware would play, what unassisted engines would play and even what skilled centaurs would play. In 2012 I proved it by attaching a password protected rar file with the moves I predicted my opponent would play, and he did all but 2 of them, and this was me 7 years ago, I have only gotten better
What were the positions for those two moves, what were the actual moves, and what were your predictions?
What matters is whether the actual moves were stronger than your predictions. (Did you win the game?)

User avatar
Ovyron
Posts: 3833
Joined: Tue Jul 03, 2007 2:30 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by Ovyron » Fri Feb 28, 2020 3:35 am

jp wrote:
Fri Feb 28, 2020 2:55 am
This may be a good test of Ovyron's wild claim he can magically tag all non-winning lines.
I'd wrongly tag Uri's position as drawn, and avoid it, and find an easier win on a previous position. From the defending side if this is a possibility then my opponent has already won and allowing this position or not makes no difference. It's a position to be avoided no matter what side I'm playing. What you'd need to prove me wrong is a position where from the middle game, the only way to win is a line that leads to this position where I can't find the single winning line, and then my magic fails, but when a position is won, I only need to find ONE win that works, and play it. Unlike the fastest mate thing where my line that mates any defense isn't enough in this artificial scenario that Zenmastur built.
jp wrote:
Fri Feb 28, 2020 2:55 am
What were the positions for those two moves, what were the actual moves, and what were your predictions?
What matters is whether the actual moves were stronger than your predictions. (Did you win the game?)
The actual moves were stronger than my predictions, but if I didn't include them on the line I'd have guessed all moves from opponent. I linked to the game in question. I was on the defending side and could only get a draw, and even in 2019 being able to predict my opponent choices wasn't enough to defeat them due to poor opening choices from my part where my opponents had easy choices. The point is that if in a Zenmastur Vs. Ovyron game I'm able to predict all his moves except for 3, but those 3 don't have an impact on the game (mating faster never has an impact on the game) then my hardware suffices, and he could do what I do without needing his big hardware.

jp
Posts: 1254
Joined: Mon Apr 23, 2018 5:54 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by jp » Fri Feb 28, 2020 3:55 am

Ovyron wrote:
Fri Feb 28, 2020 3:35 am
The point is that if in a Zenmastur Vs. Ovyron game I'm able to predict all his moves except for 3, but those 3 don't have an impact on the game then my hardware suffices
But that's what you don't know. You don't know that the 3 moves aren't superior moves that decide the game in his favor.

Just like you don't know that before you reached Uri's test position, there would be an easier win to find. That could hypothetically have been a position in the only winning line from the game.

You keep assuming that what you can't do does not matter because there would be easier choices earlier. You can't assume that.

User avatar
Ovyron
Posts: 3833
Joined: Tue Jul 03, 2007 2:30 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by Ovyron » Fri Feb 28, 2020 5:28 am

jp wrote:
Fri Feb 28, 2020 3:55 am
You can't assume that.
That's why games are played, then we know.

Like this game, before it mmt believed he could keep Leela's evals below -1.60, I assumed I could push the evals to -1.60 or worse. I didn't even know if I could win (back then I still believed 1.g4 was drawn) and just wanted to see how my attack failed (as I would have learned a lot about mmt's drawing line), but there was absolutely no way to know this in advance, unless a game is played. It has been becoming increasingly more incredible if he manages to draw, specially since his evals are -1.00 worse than mine (it'd be worrying if his evals were -1.00 better than mine, then there's a defense I'm missing, but there's attacks I haven't seen yet that make his position worse.)

Because even if Zenmastur is right, and a line exists that I can't magically tag, he'd still need to play into it in a game. Perhaps I can just avoid that line so it doesn't matter that it exists

User avatar
Ovyron
Posts: 3833
Joined: Tue Jul 03, 2007 2:30 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by Ovyron » Fri Feb 28, 2020 5:59 am

Conditional accepted.

1. g4 d5 2. Bg2 Bxg4 3. c4 c6 4. Qb3 e6 5. Qxb7 Nd7 6. Nc3 Ne7 7. cxd5 exd5 8. d4 Rb8 9. Qa6 Rb6 10. Qd3 Ng6 11. h3 Be6 12. Nf3 Bd6 13. h4 h5 14. b3 Nf6 15. Bg5 O-O 16. e3 Re8 17. Kf1 Bg4 18. Ne1 Bb4 19. Na4 Rb8 20. Nc2 Be7 21. f3 Be6 22. Nc5 Bc8 23. Kf2 Nd7 24. Ne6 Qa5 25. Bxe7 Rxe7 26. b4 Qb6 27. Ng5 Ba6 28. Qa3 Rbe8 29. Bf1 Bxf1 30. Raxf1 Qc7 31. Qd3 a5 32. a3 axb4 33. axb4



My turn again.

zullil
Posts: 5974
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by zullil » Fri Feb 28, 2020 10:10 am

Ovyron wrote:
Fri Feb 28, 2020 3:35 am
The point is that if in a Zenmastur Vs. Ovyron game I'm able to predict all his moves except for 3, but those 3 don't have an impact on the game (mating faster never has an impact on the game) then my hardware suffices, and he could do what I do without needing his big hardware.
Read what you wrote. Do you really believe it?

Playing chess is not only about predicting what an opponent will play. You also need to respond with a move that doesn't turn a draw into a loss, or a win into a non-win.

To me, it is completely obvious that you could both predict and respond better (enough so to improve your actual chess rating) if you were willing/able to upgrade your hardware. But I understand that you believe otherwise.

Uri Blass
Posts: 8730
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by Uri Blass » Fri Feb 28, 2020 10:57 am

Ovyron wrote:
Fri Feb 28, 2020 3:35 am
jp wrote:
Fri Feb 28, 2020 2:55 am
This may be a good test of Ovyron's wild claim he can magically tag all non-winning lines.
I'd wrongly tag Uri's position as drawn, and avoid it, and find an easier win on a previous position. From the defending side if this is a possibility then my opponent has already won and allowing this position or not makes no difference. It's a position to be avoided no matter what side I'm playing. What you'd need to prove me wrong is a position where from the middle game, the only way to win is a line that leads to this position where I can't find the single winning line, and then my magic fails, but when a position is won, I only need to find ONE win that works, and play it. Unlike the fastest mate thing where my line that mates any defense isn't enough in this artificial scenario that Zenmastur built.
jp wrote:
Fri Feb 28, 2020 2:55 am
What were the positions for those two moves, what were the actual moves, and what were your predictions?
What matters is whether the actual moves were stronger than your predictions. (Did you win the game?)
The actual moves were stronger than my predictions, but if I didn't include them on the line I'd have guessed all moves from opponent. I linked to the game in question. I was on the defending side and could only get a draw, and even in 2019 being able to predict my opponent choices wasn't enough to defeat them due to poor opening choices from my part where my opponents had easy choices. The point is that if in a Zenmastur Vs. Ovyron game I'm able to predict all his moves except for 3, but those 3 don't have an impact on the game (mating faster never has an impact on the game) then my hardware suffices, and he could do what I do without needing his big hardware.
one game does not prove that your hardware is good enough.

I also believe that the opening position and the position after 1.g4 are relatively easy.
It may be too easy for black to win against 1.g4 and too easy for black to draw against other moves so hardware advantage is not going to help your opponent.

It does not mean that it is so easy to get optimal results in different lines.

Zenmastur
Posts: 782
Joined: Sat May 31, 2014 6:28 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by Zenmastur » Fri Feb 28, 2020 1:04 pm

Uri Blass wrote:
Fri Feb 28, 2020 10:57 am
one game does not prove that your hardware is good enough.

I also believe that the opening position and the position after 1.g4 are relatively easy.
It may be too easy for black to win against 1.g4 and too easy for black to draw against other moves so hardware advantage is not going to help your opponent.
His strategy is to play for simply positions with as few lines of “reasonable" play as possible. This way his hardware is less of an obstacle. An example is the 1.g4 lines of play.

His opponents, if they know his weakness, should play for the most complicated lines of play possible. EVEN if it means playing less than best move. This rules out 1.g4 because you are already at such a disadvantage that it restricts your ability to make less than "perfect" moves as doing so results in an immediate loss.

In a “normal” game you would want to avoid trades, keeping the positions open, and tactically VERY sharp! This will maximize the number of lines he has to examine and therefore expose his hardware as a weakness. It's difficult to do that in 1.g4 lines. That is one of the reasons he picked that line of play. Another reason 1.g4 is so appealing to him, if he is playing black, is that the game is already decided. AB engines are very good at dealing with positions where one side has a clear advantage. This nullifies any hardware advantage a prospective opponent has against him. No amount of hardware can make up for the crippled position white has to defend.
Uri Blass wrote:
Fri Feb 28, 2020 10:57 am
It does not mean that it is so easy to get optimal results in different lines.
Precisely!

Regards,

Zenmastur
Only 2 defining forces have ever offered to die for you.....Jesus Christ and the American Soldier. One died for your soul, the other for your freedom.

SheikhYerbouti
Posts: 18
Joined: Mon Jan 27, 2020 3:25 pm
Full name: Daniel Kemmer

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by SheikhYerbouti » Fri Feb 28, 2020 3:17 pm

Zenmastur wrote:
Fri Feb 28, 2020 1:04 pm


His strategy is to play for simply positions with as few lines of “reasonable" play as possible. This way his hardware is less of an obstacle. An example is the 1.g4 lines of play.

His opponents, if they know his weakness, should play for the most complicated lines of play possible. EVEN if it means playing less than best move. This rules out 1.g4 because you are already at such a disadvantage that it restricts your ability to make less than "perfect" moves as doing so results in an immediate loss.

In a “normal” game you would want to avoid trades, keeping the positions open, and tactically VERY sharp! This will maximize the number of lines he has to examine and therefore expose his hardware as a weakness. It's difficult to do that in 1.g4 lines. That is one of the reasons he picked that line of play. Another reason 1.g4 is so appealing to him, if he is playing black, is that the game is already decided. AB engines are very good at dealing with positions where one side has a clear advantage. This nullifies any hardware advantage a prospective opponent has against him. No amount of hardware can make up for the crippled position white has to defend.

Precisely!

Regards,

Zenmastur
Indeed, this has wider implications for chess in general! In addition, I edited out Uri's posts to save space but also make your efforts self-congratulatory. We know that Stockfish self-play after g4 results in a draw given sufficient nodes. We can also surmise g4 is worst opening move for white. Therefore with perfect play white will at least hold and black should also hold. After all if g4 holds for white then surely every 'perfect' counter move for black will also hold, after whatever first move white plays

it does show Stockfish as usual performing very well with sharp tactical openings and Leela less well. Of course Leela doesn't favor imbalanced positions full stop. Further evidence to this is her seeming endless path to checkmate. Like she's trying not to lose rather than win. So why not promote a pawn to a knight... to prolong it for fun?

Post Reply