TalkChess.com

If SF really thought it was winning you'd see much higher scores. The oscillation in score suggests it's not really sure - it might be winning, but then again it might not.

Then why is the Leela site stating this? "As our previous 'run to completion' has not made much progress recently , main page and default contributions have been reset to our new actually testing run while we wait for lc0 0.20.0 to be prepared with a new network architecture for our next main run." ...

To avoid needlessly chasing its own proverbial tail, I still think they should also test LC0 against known entities with a fixed, established rating - at least periodically. That should automatically be part of the testing protocol.

2...d6 is also good and may well transpose into lines with Nc6
2...e5 is not that good for reasons given by a prior poster

I raised a similar point. Yes, it can achieve a winning endgame in grandmasterly fashion, but just when things should be super easy they become almost impossibly hard, taking the longest possible time to win such a simple ending. Could be a faulty TB implementation? And why are the depths getting lo...

Take a look at this thread,

http://talkchess.com/forum3/viewtopic.php?f=2&t=68963

for a possible explanation.

If self-play is totally broken and meaningless, why don't they test against engines of established known strength?

tpoppins wrote: ↑

Thu Nov 22, 2018 8:16 pm

lech wrote: ↑

Thu Nov 22, 2018 1:33 pm

The authors of studies selected to the "IQ" test:

by YACPDB:
<...>
Klyackin Mihail - 65
<...>

That should be Klyatskin Mikhail, cf. CN 646.

(He was merely using the Polish spelling.)

Thanks for the updated Sting!
CL

Where can network 11248 be downloaded?
Thanks!

I don't understand why Leela isn't being trained intensively on simple endgames as part of its training. Today I've found that the Leela opencl (network=31269, apparently very recent) running on a GPU has serious trouble winning the most basic K+Q vs Q ending. It plays in a completely aimless manner...