Search found 37 matches

by megamau
Sun Aug 12, 2018 7:25 am
Forum: Computer Chess Club: General Topics
Topic: Do you understand the following graph lc0-stockfish?
Replies: 4
Views: 965

Re: Do you understand the following graph lc0-stockfish?

These are *tiny* number of nodes. 1 or 10 nodes per move translate in a timecontrol in the region of milliseconds. Also in these first part of the curve, the nodes per second vary quite a lot (as identified by Kai). Stockfish indeed "scales" better from 1k nodes/move to 1M/move with respect of Leela...
by megamau
Fri Aug 10, 2018 12:58 pm
Forum: Computer Chess Club: General Topics
Topic: Price comparison of CPU and GPU setup
Replies: 1
Views: 440

Price comparison of CPU and GPU setup

The discussion in another thread has spurred me to prepare a table. The one below is only the skeleton, but I would need the help of the forum to fill it up (and actually also to format it :-) ) What is the best setup for given price and what is the NPS ratio ? How does this translate to Elo differe...
by megamau
Mon Jul 30, 2018 11:59 pm
Forum: Computer Chess Club: General Topics
Topic: Something goes wrong with lc0 since yesterday?
Replies: 272
Views: 45860

Re: Something goes wrong with lc0 since yesterday?

Is Leela still making progress ? I ask because all new id are still loosing against id 390. The graph seems not reliable at all. What do you mean ? All the latest tests ID390 has lost. http://lczero.org/matches Id Candidate ID Current ID Pass Score Elo Delta Elo Error Margin Time 549 390 527 test +...
by megamau
Wed Jul 25, 2018 2:10 pm
Forum: Computer Chess Club: General Topics
Topic: Scorpio 2.8
Replies: 35
Views: 10938

Re: Scorpio 2.8

Daniel Shawul wrote:
Wed Mar 28, 2018 3:49 pm
My interal measurements puts it about 2300 ccrl (nebiyu/joker level), which is about 600 elo above TSCP and about 800 elo above current leela-chess
Good luck catching upto that leela-chess fans
Have you tried testing lately ?
I think the Leela-Chess fan had quite a good luck in catching up.
by megamau
Sat Jul 07, 2018 7:25 pm
Forum: Computer Chess Club: General Topics
Topic: Stockfish's Asymmetric Evaluations
Replies: 15
Views: 2223

Re: Stockfish's Asymmetric Evaluations

I guess that stockfish evaluates a draw by repetition as 0.00 regardless of contempt. That doesn't really make sense to me. With a positive contempt, don't you want draws, especially forced draws such as draws by repetition often are, to be scored lower than zero? Otherwise you have no incentive to...
by megamau
Thu Jul 05, 2018 4:49 pm
Forum: Computer Chess Club: General Topics
Topic: Something goes wrong with lc0 since yesterday?
Replies: 272
Views: 45860

Re: Something goes wrong with lc0 since yesterday?

yanquis1972 wrote:
Thu Jul 05, 2018 4:13 pm
has whatever gating policy leelazero has in place ever been acted upon?? or is it constantly happening, & the labels either misleading or i'm not reading them correctly? it seems like the candidate id always gets promoted
LeelaZero is the GO program.
http://zero.sjeng.org/
by megamau
Thu Jul 05, 2018 3:30 pm
Forum: Computer Chess Club: General Topics
Topic: Something goes wrong with lc0 since yesterday?
Replies: 272
Views: 45860

Re: Something goes wrong with lc0 since yesterday?

Their self-plays by now are almost completely useless. If I am not doing something very wrong, or my time control is too short, I am getting for ID10030 almost identical result to ID10028 in 200 games gauntlets against AB engine. Still only some 50-60 Elo points improvement compared to ID10017, whi...
by megamau
Mon Jun 25, 2018 5:36 pm
Forum: Computer Chess Club: General Topics
Topic: Something goes wrong with lc0 since yesterday?
Replies: 272
Views: 45860

Re: Something goes wrong with lc0 since yesterday?

Well, what I dont understand is that the original branch had i.e. ID 107 with 6x64 at 4780 which is much higher/stronger right? rgds I doubt it. The elo of self play is very distorted so it's hard to know what it means without independent testing like Kai does. I also do not understand how they hav...
by megamau
Mon Jun 25, 2018 1:52 pm
Forum: Computer Chess Club: General Topics
Topic: LC0 finally?
Replies: 21
Views: 5529

Re: LC0 finally?

It is actually quite simple. It show how different nets (column C) behave with time odds against Stockfish 9. Column D shows the results when both Leela and Stockfish use only 1 node per move. Column E shows the results when they both search 10 nodes per moves etc. The green columns use same number ...
by megamau
Fri Apr 06, 2018 12:23 pm
Forum: Computer Chess Club: General Topics
Topic: LCZero: Progress and Scaling. Relation to CCRL Elo
Replies: 604
Views: 108299

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

Only after some 1 million playouts it switches to the winning Kf4. I don't know how these MCTS rollouts work here, but my experience with MCTS Go engines, especially the newer Leela or Crazy Stone is different. Tactically this blunder (from won position to -M4) is equivalent to losing an elementary...