mmt Vs. Ovyron (G4 D5 BG2)

[Ovyron]

Bring in the best you can Jim!

[Zenmastur]

A little update: I'm changing LC0's weights to net 59866. I haven't seen or done enough testing to know that my previous bigger Sergio's net is better. I'm also updating the parameters for wider search because I'm doing very long calculations and I'm using SF to guide LC0 if there are tactical problems. I will be recompiling LC0 to use doubles instead of floats like koedem has suggested - this seems more important in long searches.

h4 is likely next.

I'm curious how deep your SF searches are vice your Lco searches.

Regards,

Zenmastur

[mmt]

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

[d]3qk2r/p2n1ppp/1rpbb1n1/3p4/3P3P/2NQ1N2/PP2PPB1/R1B1K2R b KQk - 0 13
I'm very interested in the response. I've stayed with a larger net based on LC0 Discord recommendations.

[mmt]

I'm curious how deep your SF searches are vice your Lco searches.

I'm using Nibbler for LC0 lately and it doesn't show depth. I'm usually just verifying various LC0 lines in SF but if I run a full long search, it gets to around 51.

[zullil]

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

[d]3qk2r/p2n1ppp/1rpbb1n1/3p4/3P3P/2NQ1N2/PP2PPB1/R1B1K2R b KQk - 0 13
I'm very interested in the response. I've stayed with a larger net based on LC0 Discord recommendations.

Komodo 13.2.5 at depth 45 using MultiPV=3: Black's three best options have evaluations of -1.19, -1.12 and -1.11. Won't give the associated moves!

[mmt]

At depth 44, Komodo has -1.09. I'll refrain from giving its current favorite for White's move 13, since White hasn't played that move yet.

So was that mystery move h4?

[zullil]

At depth 44, Komodo has -1.09. I'll refrain from giving its current favorite for White's move 13, since White hasn't played that move yet.

So was that mystery move h4?

Yes.

[jp]

Komodo 13.2.5 at depth 45 using MultiPV=3: Black's three best options have evaluations of -1.19, -1.12 and -1.11.

This is where it gets interesting -- when move evals are very close. Generally, it could mean the engine doesn't know what to do (or it could mean the moves really are much the same).

[zullil]

Komodo 13.2.5 at depth 45 using MultiPV=3: Black's three best options have evaluations of -1.19, -1.12 and -1.11.

This is where it gets interesting -- when move evals are very close. Generally, it could mean the engine doesn't know what to do (or it could mean the moves really are much the same).

Yes. And I didn't check if the three PVs include transpositions and might converge to a common future position ...

[mmt]

Yes. And I didn't check if the three PVs include transpositions and might converge to a common future position ...

From what I've seen, the lines in this game do not converge (order matters a lot).