Most Recent
Forum Legend
Following
New Comments
Locked Topic
Pinned Topic
I've seen some articles assessing the bots' ratings and whether they are accurate, but I haven't seen any discussing HOW the ratings are assigned. Does anyone know what their ratings are based on? Do they beta test these things with a bunch of human players of various strengths? Do they have them play against different versions of established engines like stockfish or komodo with known elos?