Dylan Sharp Vs. Harvey Williamson (G4)

[Ovyron]

It has 'gotten' boring ....hopefully it picks up...

Hopefully not. What do you want, black sacrificing its queen? The variation is there but white loses, so I'd be crazy if I allowed Harvey to play it.

There's a ton of possible fun variations and material imbalances that could happen, but all of them are either red flags (black has a move that makes the tree of variations and branching factor explode, so I can't even check other black alternatives) or white loses, so my job is to keep things "boring." Harvey's job would be to make things unclear, because it's white's job to clarify them, while black only needs a line up to a big advantage.

The game is about being able to tell the drawn positions from the losing positions apart and avoid the latter ones. So far losing positions behave clearly differently, so Harvey needs to make me play into a lost position that behaves like drawn ones.

[jp]

I guess it's hard to be motivated when you have a lost position so early.

It's only -1.64 at depth 72. No reason to give up hope. <snip>

And also -1.64 at depth 73. The current PV stretches to Black's 67th move.

Yes. SF hasn't found a win yet, and the eval is not really increasing yet, so there could be a draw.

[Ovyron]

I want to thank Harvey for being a good sport. His move ordering and conditionals have been saving me a lot of time (for instance, all the analysis of 14...Re8 and 14...Rc8 that didn't need to be expanded after knowing he'll play 14...Qd7.)

1. g4 d5 2. g5 e5 3. d4 exd4 4. Nf3 c5 5. Bg2 Ne7 6. c3 dxc3 7. Nxc3 Nbc6 8. O-O d4 9. Ne4 Ng6 10. a3 Bg4 11. Bd2 Bd6 12. Rc1 O-O 13. Nxc5 Bxc5

[d]r2q1rk1/pp3ppp/2n3n1/2b3P1/3p2b1/P4N2/1P1BPPBP/2RQ1RK1 w - -

My turn again.

[zullil]

I want to thank Harvey for being a good sport. His move ordering and conditionals have been saving me a lot of time (for instance, all the analysis of 14...Re8 and 14...Rc8 that didn't need to be expanded after knowing he'll play 14...Qd7.)

1. g4 d5 2. g5 e5 3. d4 exd4 4. Nf3 c5 5. Bg2 Ne7 6. c3 dxc3 7. Nxc3 Nbc6 8. O-O d4 9. Ne4 Ng6 10. a3 Bg4 11. Bd2 Bd6 12. Rc1 O-O 13. Nxc5 Bxc5

[d]r2q1rk1/pp3ppp/2n3n1/2b3P1/3p2b1/P4N2/1P1BPPBP/2RQ1RK1 w - -

My turn again.

OK, I can now safely say that Stockfish-dev had 11...Be7, rather than 11...Bd6. But of course, they both lead to the same place with Bxc5. So we're back to the same PV that Stockfish had after 2. g5, which it has kept essentially unchanged through more than a thousand search iterations since then.

Still awaiting a "novelty".

[Ovyron]

Still waiting for a 20 core machine to show any use for its resources

[zullil]

Still waiting for a 20 core machine to show any use for its resources

It's warming my unheated basement.

[jp]

It's warming my unheated basement.

If massive heating is your goal, I suggest unleashing Leela with the GPUs. Maybe wait for winter for that.

[Ovyron]

Leela is something I can't emulate, so it'd be much more interesting to get its evals.

[zullil]

Leela is something I can't emulate, so it'd be much more interesting to get its evals.

I could run both, but I only have 128 GB RAM total, and I've already given 64 GB to Stockfish.

I assume someone else can provide Lc0 evaluations.

[Ovyron]

Haha, I only got 3GB of RAM yet I can run CPU Leela and Stockfish in parallel. Talk about first world problems XD