https://docs.google.com/spreadsheets/d/ ... edit#gid=0
and
https://docs.google.com/spreadsheets/d/ ... edit#gid=0
1st spr. look the id307 and 2nd spr. the id305
lc0 elo: more confusion
Moderators: hgm, Rebel, chrisw
-
- Posts: 6340
- Joined: Mon Mar 13, 2006 2:34 pm
- Location: Acworth, GA
Re: lc0 elo: more confusion
Isn't one versus various versions of Stockfish only, and the other versus various Engines in general?
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
__________________________________________________________________
Ted Summers
-
- Posts: 1766
- Joined: Wed Jun 03, 2009 12:14 am
Re: lc0 elo: more confusion
i thought the other was CCRL, but i can't find it. i was going to start a test like this
anyway, the 200games list is shockingly not-at-all at odds with self-play ELO; it actually shows bigger gains since the last apparent significant change was. also suggests lc0 has finally caught or surpassed pre-fuckup overall elo.
from Lc0 builds 288 - 307 self-play/Lc0 v SF ELO = +218/+219
between almost identical gain & this being the only guy i've seen publishing regular & consistent tests on reasonable hardare, it looks like a good companion to the self play chart. doubt we'll see anybody stop mocking it though.
anyway, the 200games list is shockingly not-at-all at odds with self-play ELO; it actually shows bigger gains since the last apparent significant change was. also suggests lc0 has finally caught or surpassed pre-fuckup overall elo.
from Lc0 builds 288 - 307 self-play/Lc0 v SF ELO = +218/+219
between almost identical gain & this being the only guy i've seen publishing regular & consistent tests on reasonable hardare, it looks like a good companion to the self play chart. doubt we'll see anybody stop mocking it though.