Search found 42 matches

by mirek
Fri Apr 20, 2018 12:06 am
Forum: Computer Chess Club: General Topics
Topic: LCzero sacs a knight for nothing
Replies: 106
Views: 16376

Re: LCzero sacs a knight for nothing

That must be due to luck then that ID149 solved it, ID150 could go back to f3e5 move and what will you then say it learned? Must be? "may be" would be more appropriate. And obviously I was trolling you a bit there, but OK, let's make collection of shallow tactical blunders of ID 125 from TCEC and s...
by mirek
Thu Apr 19, 2018 8:47 pm
Forum: Computer Chess Club: General Topics
Topic: LCzero sacs a knight for nothing
Replies: 106
Views: 16376

Re: LCzero sacs a knight for nothing

Here LC0 moved Ne5 on TCEC's 43-core hardware! Note that this blunder is probably not due to a bug as most other engines would have it, but that the algorithm is working as intended and can produce such tactical blunders even on this massive hardware. Are you telling me that this is not a problem f...
by mirek
Wed Apr 18, 2018 9:40 pm
Forum: Computer Chess Club: General Topics
Topic: LC0 on 43 cores had a ~2700 CCRL ELO performance.
Replies: 58
Views: 8540

Re: LC0 on 43 cores had a ~2700 CCRL ELO performance.

If you use MCTS alone, you will suffer from tactical problems even a 100x more time won't solve a 7-ply trap. If we are speaking A0 that is only true if the search guiding NN won't recognize the patterns and realize that such trap maybe there. And obviously NN can fail at recognizing such pattern s...
by mirek
Wed Apr 18, 2018 9:21 pm
Forum: Computer Chess Club: General Topics
Topic: LC0 on 43 cores had a ~2700 CCRL ELO performance.
Replies: 58
Views: 8540

Re: LC0 on 43 cores had a ~2700 CCRL ELO performance.

They had no option but to use MCTS not because it is better. That is because it was getting 80,000 nodes/s even on 4-TPUs. With that nps a full width alpha-beta search you are stuck with search depth engines used to get in the 90's. That brings up tactical problems which they minimized with massive...
by mirek
Wed Apr 18, 2018 7:30 pm
Forum: Computer Chess Club: General Topics
Topic: LC0 on 43 cores had a ~2700 CCRL ELO performance.
Replies: 58
Views: 8540

Re: LC0 on 43 cores had a ~2700 CCRL ELO performance.

c) Hardware differences. It has become very clear to me this result is achieved via massive hardware acceleration of a very slow eval. Theoretically Deep Blue could also have acieved this result with their FGPA. Admittedly, their approach is cost-effective given the future is cheap manycore archite...
by mirek
Wed Apr 18, 2018 1:33 pm
Forum: Computer Chess Club: General Topics
Topic: The real elo of lczero is in its name
Replies: 73
Views: 12547

Re: Leela 43 cores Versus Scorpio 32 cores at TCEC

Then you have the issues with the network. ID 125 is one of the first networks to come out after v0.5 was introduced, so it was mostly trained with bug-free matches. After that, the number of bugged matches started to increase eventually filling the training window with 100% bugged matches, that's ...
by mirek
Wed Apr 18, 2018 9:00 am
Forum: Computer Chess Club: General Topics
Topic: LC0 on 43 cores had a ~2700 CCRL ELO performance.
Replies: 58
Views: 8540

Re: LC0 on 43 cores had a ~2700 CCRL ELO performance.

What Daniel is constantly pointing is that alpha-beta as general search might not have the future, but there is absolutely zero evidence that NN evaluation is the key. A0 got its performance from massive amount of rollouts thanks to massive hardware. If you used better MCTS and AB for leaves evalua...
by mirek
Wed Apr 18, 2018 1:22 am
Forum: Computer Chess Club: General Topics
Topic: LC0 on 43 cores had a ~2700 CCRL ELO performance.
Replies: 58
Views: 8540

Re: LC0 on 43 cores had a ~2700 CCRL ELO performance.

But this LC0 engine could actually be a 3500 elo on 4 TPUs, that just seems hard to attribute to algorithmic breakthrough ... I think this greatly depends on time control. If for example we were speaking 1 min / move as in A0 paper we could estimate expected performance of A0 on single high-end con...
by mirek
Tue Apr 17, 2018 9:16 am
Forum: Computer Chess Club: General Topics
Topic: LCZero: Progress and Scaling. Relation to CCRL Elo
Replies: 604
Views: 107539

Re: LCZero: Progress and Scaling. Relation to CCRL Elo

Here is a recent result of Carlos Canavessi till ID138, it also shows a significant regression. Oh no, any explanation why is this happening? Is it already a reason for concern or is it still OK? And at which point if the trend continues will it become reason for concern? Shouldn't have the larger ...
by mirek
Sun Apr 15, 2018 6:12 am
Forum: Computer Chess Club: General Topics
Topic: LCZero tactical improvement
Replies: 9
Views: 1414

Re: LCZero tactical improvement

Sorry but what kind of "knowledge" is that when you have to perform at least 10k searches per second, i.e. probe 10k lines to reach 3000 Elo level??? Are you referring to the strength scaling graph in A0 paper? If so I guess that numbers you get depend on what rating you assign to Stockfish in that...