| General Chess Discussion - Introducing Maia, a human-like neural network chess engine#80 @FC-in-the-UK Thanks for clarifying. I don't think we should read too much into the shape of the curves. We are using Stockfish, plus a rough approximation of win probability to judge each position so… | reidmcy |
| General Chess Discussion - Introducing Maia, a human-like neural network chess engine#76 @mars69ha, @TheRisottoVariation The maia bots are only accepting timed games. If the opponent leaves during a Correspondence game it causes issues for the backend. @Eyvazova_2009, @jrepo Thanks for th… | reidmcy |
| Lichess Feedback - maia9 has a bug#14 I'm the main dev on Maia, thanks for mentioning this @RISHABHA . This is an issue with all the maia bots, since they use a single deterministic model with no randomness added. I'm working on adding ra… | reidmcy |
| Game analysis - Finding interesting boards computationally #2 The rules I'm using right now for a board to be interesting: + Has at least 2 lines + Exactly one line has a difference in rating between the best and second best move above 1 + The last move of all l… | reidmcy |
| Game analysis - Finding interesting boards computationally #1 I've been trying to find boars that lead to interesting decisions automatically. Right now I'm looking for boards where there's a 'tricky' line and a 'safe' line. I found some using a combination of t… | reidmcy |