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.
Zenmastur
Posts: 846
Joined: Sat May 31, 2014 6:28 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by Zenmastur » Sun Feb 02, 2020 1:07 am

zullil wrote:
Sat Feb 01, 2020 11:14 pm
mmt wrote:
Sat Feb 01, 2020 9:44 pm
1. g4 d5 2. c4 Bxg4 3. Bg2 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


This was really hard to choose. LC0 by itself prefers 14. Ng5 (-1.27) but when pushed further along the Ng5 line it weakens to <-1.40. When pushed ahead with 14. b3, it gets the exact same score -1.27. Many moves are semi-forced with Ng5. SF 11 sees trouble ahead for both of these options with LC0's preferred responses (around -1.50 to -1.65). Now I'm quite sure that SF's eval is better in this position because of how deep it can see.
Komodo at depth 47 with MultiPV = 3 had

14. b3 (-1.06)
14. Ng5 (-1.10)
14. Bh3 (-1.32)

Interestingly, Komodo had Ng5 as its first choice until depth 46.
Hmmm... I searched this with SF11 Jan 20 version to depth 59 (1.6 Trillion nodes, score -1.61) and then did a reverse analysis (average depth ~60, greatest depth 67) on the given line, which started with 14.b3. I ended up with a score of -1.83 for what it's worth. Based on what you've been posting about Komodo's analysis I wouldn't/don't trust it even if it's done to great depth. Just my opinion!

One other note, when using Multi-PV on long runs (i.e. any length of time in which the number of nodes searched is greater than 10 times the number of TT entries) you should ALWAYS verify the best line with multi-PV off. I let you figure out why this is the case.

I did the same type of analysis on blacks 13th move. It gave 13. ... h5 as best.

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.

zullil
Posts: 6157
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 » Sun Feb 02, 2020 2:11 am

Zenmastur wrote:
Sun Feb 02, 2020 1:07 am
zullil wrote:
Sat Feb 01, 2020 11:14 pm
mmt wrote:
Sat Feb 01, 2020 9:44 pm
1. g4 d5 2. c4 Bxg4 3. Bg2 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


This was really hard to choose. LC0 by itself prefers 14. Ng5 (-1.27) but when pushed further along the Ng5 line it weakens to <-1.40. When pushed ahead with 14. b3, it gets the exact same score -1.27. Many moves are semi-forced with Ng5. SF 11 sees trouble ahead for both of these options with LC0's preferred responses (around -1.50 to -1.65). Now I'm quite sure that SF's eval is better in this position because of how deep it can see.
Komodo at depth 47 with MultiPV = 3 had

14. b3 (-1.06)
14. Ng5 (-1.10)
14. Bh3 (-1.32)

Interestingly, Komodo had Ng5 as its first choice until depth 46.
Hmmm... I searched this with SF11 Jan 20 version to depth 59 (1.6 Trillion nodes, score -1.61) and then did a reverse analysis (average depth ~60, greatest depth 67) on the given line, which started with 14.b3. I ended up with a score of -1.83 for what it's worth. Based on what you've been posting about Komodo's analysis I wouldn't/don't trust it even if it's done to great depth. Just my opinion!

One other note, when using Multi-PV on long runs (i.e. any length of time in which the number of nodes searched is greater than 10 times the number of TT entries) you should ALWAYS verify the best line with multi-PV off. I let you figure out why this is the case.

I did the same type of analysis on blacks 13th move. It gave 13. ... h5 as best.

Regards,

Zenmastur
Thanks for Stockfish's opinion. I'm running Komodo simply because it's not Stockfish, to provide some diversity. I'm not providing any guidance to the engine, or doing any forward/backward analysis. No time for that.

mmt
Posts: 216
Joined: Sun Aug 25, 2019 6:33 am
Full name: .

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by mmt » Sun Feb 02, 2020 2:50 am

Zenmastur wrote:
Sun Feb 02, 2020 1:07 am

Hmmm... I searched this with SF11 Jan 20 version to depth 59 (1.6 Trillion nodes, score -1.61) and then did a reverse analysis (average depth ~60, greatest depth 67) on the given line, which started with 14.b3. I ended up with a score of -1.83 for what it's worth. Based on what you've been posting about Komodo's analysis I wouldn't/don't trust it even if it's done to great depth. Just my opinion!
Yes, around -1.5 to -1.7 is the score that I saw for SF for both possibilities (b3 and Ng5). SF can diverge from LC0 quite a bit in the next moves but LC0 doesn't go deep enough to see it get worse and it needs to be pushed.

User avatar
Ovyron
Posts: 4171
Joined: Tue Jul 03, 2007 2:30 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by Ovyron » Sun Feb 02, 2020 3:36 am

What I want to know is if this game looks better for white than my game against Harvey on move 14, or if I was just doomed at this point and was just very oblivious about it. In both games black's best line was almost forced (unless moves like 13...O-O win as well) but being able to predict what black will play isn't helping white draw the game [unless I don't win].

Zenmastur
Posts: 846
Joined: Sat May 31, 2014 6:28 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by Zenmastur » Sun Feb 02, 2020 5:34 am

Ovyron wrote:
Sun Feb 02, 2020 3:36 am
What I want to know is if this game looks better for white than my game against Harvey on move 14, or if I was just doomed at this point and was just very oblivious about it. In both games black's best line was almost forced (unless moves like 13...O-O win as well) but being able to predict what black will play isn't helping white draw the game [unless I don't win].
You can't look deep enough with your hardware. Your TT size limits how deep you can probe and get accurate answers. I was using 96 GB of TT and that was almost too small. What I would really like is to have a full 1TB of TT. With 96GB between 45 and 55 plies tends to fill it up and then the number of nodes that need to processed goes way up, so the search starts slowing way down. This limits how deep you can search in a reasonable time. But, I must say this system makes it possible to analyze positions in a few hours that would have taken several day on my old system. The only down side is the amount of heat it generates.

A sure sign that your in a bad way is the score gets worse the deeper you search. You can see this even on a slow machine if you uses compound searches. But you need more TT to be able to do that effectively.

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.

mmt
Posts: 216
Joined: Sun Aug 25, 2019 6:33 am
Full name: .

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by mmt » Sun Feb 02, 2020 7:19 am

Ovyron wrote:
Sun Feb 02, 2020 3:36 am
What I want to know is if this game looks better for white than my game against Harvey on move 14, or if I was just doomed at this point and was just very oblivious about it. In both games black's best line was almost forced (unless moves like 13...O-O win as well) but being able to predict what black will play isn't helping white draw the game [unless I don't win].
Link to your previous game?

User avatar
Ovyron
Posts: 4171
Joined: Tue Jul 03, 2007 2:30 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by Ovyron » Sun Feb 02, 2020 7:53 am

I have never needed more than 128 MB for the TT, and never have needed high depth to analyze.
mmt wrote:
Sun Feb 02, 2020 7:19 am
Link to your previous game?
It's here.

And my move:

1. g4 d5 2. c4 Bxg4 3. Bg2 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


User avatar
Ovyron
Posts: 4171
Joined: Tue Jul 03, 2007 2:30 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by Ovyron » Sun Feb 02, 2020 8:17 am

Oh, here's the full game without the forum messages in-between (courtesy of Uri Blass):


Zenmastur
Posts: 846
Joined: Sat May 31, 2014 6:28 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by Zenmastur » Sun Feb 02, 2020 8:28 am

Ovyron wrote:
Sun Feb 02, 2020 7:53 am
I have never needed more than 128 MB for the TT, and never have needed high depth to analyze.
Is that a fact?
Ovyron wrote:
Sun Feb 02, 2020 8:17 am
Oh, here's the full game without the forum messages in-between (courtesy of Uri Blass):

Then why did this happen? It certainly looks to me like you need both more TT and a faster processor!

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.

jp
Posts: 1341
Joined: Mon Apr 23, 2018 5:54 am

Re: mmt Vs. Ovyron (G4 D5 BG2)

Post by jp » Sun Feb 02, 2020 8:56 am

Zenmastur wrote:
Sun Feb 02, 2020 8:28 am
Then why did this happen? It certainly looks to me like you need both more TT and a faster processor!
Well, Ovyron thinks it's because the opening is lost! Would you be willing to play White in the next game, if Jim doesn't succeed in avoiding defeat in this one?

Post Reply