-
**The calculation of the rating after each match can be optimised and better.**
The function for calculating the ELO-rating can be found in the `matchmaker.service.ts` named `calculateELORating ` …
-
The progress figure at http://zero.sjeng.org shows an ELO-scale with respect to random player=0. Is there any possibility to add a kyu/dan/pro scale, at least approximately? I guess this is what most …
-
I am playing with the Idea of creating a teacher using leelaz and a bunch of networks.
I would like to have one network for each rank from 30 Kyu to 5 Dan, so each player can against it own level. Th…
-
After a very long and interesting discussion https://github.com/official-stockfish/Stockfish/pull/1804#issuecomment-445429885 I think it's time to summarize in quantitative way what people have found…
-
-
https://cloudygo.com/v16-19x19/eval-model/986?sorted=False
-
We can see on zero.sjeng.org that after 265.000 games LZ is above 1200 ELO. It means about 8kyu (european). But if you look at any of the last games they clearly don't look like that strength. Can any…
-
To avoid 0-10 elo change games I propose to limit bot matches to those that can award at least 1-9 elo points. To make this more intuitive I would say we should introduce leagues that match these rang…
-
I see a regression test from 07-12-18 with only +15.12 improvement.
But I see only one improvement 3 weeks ago, worth 65 rating points !
How is it possible ?
Thanks !
Details :
07-12-18 ia…
-
`SyncGTPEngineCommand` uses the GTP command `fixed_handicap` as part of synchronizing the application's and Fuego's internal board models. This GTP command assumes that the current game's handicap sto…