Life In 19x19

For me the most interesting to see is how the bot likes (or dislikes) the normal human moves over time, and the same for the moves from the Alpha Go Zero games. Where has it started rediscovering theory, and where is it inventing new things...

The last example in this quick overview is the pattern that is #1000 in global popularity: invasion at 1x1 against the opponent's 4x4! Interestingly, this still comes up in relatively recent versions, every now and then. That happens very rarely (about once in 250K patterns, and is mostly likely ...

Obviously AG had to be given 1) the basic rules, 2) counting & scoring and 3) the goal of the game in advance. You cannot learn that as an AI. We has humans can because we can look outside of the game and learn 1)-3) from that "higher perspective", e.g. when we see multiple times that someone ...

Can someone test Zen 7 vs Leela 11 GPU version?


Leela-0.11.0-8t1g 3464 198 2017-10-24 21:49:09
Zen-15.3-1c0g 3457 4263 2017-10-25 07:32:31
Zen-15.4-1c0g 3449 922 2017-10-25 07:32:31


Zen should be stronger. Zen is only a little behind when using 1 core ...

The license in ThreadPool.h is not compatible with the GPLV3...My concern is that generally anything distributed alongside GPLv3 code must also be GPLv3.

That's simply completely wrong: m

There are a lot of open source licenses compatible with GPLv3, including the license in that file, which ...

Given the release notes, it sounds like the playouts are a lot heavier than before, with a lot more attached to them. Playout count may go down, but overall level and playing quality should increase.

From the same release notes:
"More accurate, larger value network for 19x19."

Larger network ...

The relative efficiencies of the operating systems should matter only when a significant percentage of the time is spent in them (system calls, other dynamic calls).

He's testing the GPU acceleration, so that's exactly what happens. The CPU is forwarding the work to the GPU, through the OS ...

I rechecked and here is what Leela says in detail: D2(best move): Win% 42, MC win% 51.7, Net win% 31.8, Net prob% 29. What is the meaning of all these numbers anyway? I always thought that win% matters, the rest not so much.

It's explained on Leela's webpage.

FAQ: "What do the columns in the ...

Given that you are 5D, can't you just play this out vs Leela, and see what happens after a bunch of moves? Either it'll show you a refutation, or its score will go towards 50%.

I saw somewhere (compgo?) that the Leela author said he is not running Leela on CGOS. So it is an unofficial test by someone else.

I think it's a test by pookpooi actually.

The most interesting would probably be to run Leela with 1 cpu and no GPU. Then we know how much the GPU is worth. I'm not ...

http://www.yss-aya.com/cgos/19x19/bayes.html

Leela 0.9.1 is out in the results. It appears that new leela is significantly weaker than RN3.9, The difference could be several stones.


I think you misread, it is not Leela 0.9.1, it is Leela 0.9 running on a single core. It's stronger than Rn3.8 on ...

goame wrote:Is it possiple to use 4x nvidia geforce gtx 1080 ti with Leela?

No, because you cannot buy those yet

76 ELO ahead of the commercial Zen

So Leela (with GPU) is between half a stone and a stone stronger than Zenith Go 6. johnsmith called it, Gomoto was wrong.

DGS uses Japanese rules, and Leela uses Chinese ones exclusively, so Leela won't play correctly on that server.

In any case playing random machine A against random machine B for a single game is a pretty silly way to compare two programs which you can just download. There's several tools that can ...

My interpretation is that Zen developers failed to improve after 13.3.


12.5 is actually the highest rated one. Very sad that Zen didn't improve, then, but I do wonder if it's really true.


1 Zen-12.5n-1c1g 3815 147 147 188 97% 2768 0%
2 Zen-13.1-0.5-1c0g 3799 52 52 881 95% 3069 0 ...