-
Everytime a user select an option in the test view a new ELO rating is calculated. A solution whould be to store the choice in an array and then calculating the ELO rating when the user pushes the con…
-
As of now Fishtest only tracks single core regression at LTC.
For multicore tests there is NCM dev but the problem I see is that SF dev is playing **SF7** which might not be the best opponent to c…
-
-
When will the mlb-elo.csv file be updated to include the full 2019 schedule and the season-opening elo ratings? Thanks!
-
Be able to play a 2v2 match, where the players on the winning team gets rating.
Suggestion for calculations:
The rating of the team is the average of the individual players on the team:
Team 1: 150…
-
http://www.yss-aya.com/cgos/19x19/cross/LZ_223_c9fb_p400.html
-
I wanted to test scaling with node count of a few engines to compare with SF. I'm running a cutechess tournament with a cutechess-cli option to limit nodes for an engine.
However, whereas SF result…
-
Judging by the game counts at , Elo rating of the new networks are calculated mostly from the winning percentage against the previous ones. This is not very precise, since the networks may be overfitt…
-
I did a SPRT validation for these two networks. The winning c4611a60 (4958) and the loosing 657f91c7 (4244) so the Elo rating for GnuGo would be 4564 in our chart. Gnugo is rated 5 kyu so it is 1600 …
-
So if we want to obtain reasonable Elo ratings without self-bias, it is essential to anchor ratings with, for example, GNUGo 3. Often more than one anchors are essential if the ratings are distributed…