3 ms·

Alph zero/mu at similar computation resources was weaker for a while iirc.

• joshuamorton 5y ago
  Right, my recollection is that stockfish + nnue has been stronger than leela since stockfish 13, and leela is stronger than A0.
  • cellis 5y ago
    Isn't the difference though, that Leela is a pure RL implementation whereas NNUE/SF still uses some tablebase stuff?
    • TchoBeer 5y ago
      No, NNUE doesn't necessarily need tablebases and Leela uses tablebases too. The difference is that their network architecture are different and their search algos are different.
    • ivalm 5y ago
      Leela uses mcts, sf does not. SF also trains on human positions, while I think Leela is self play only.
      • adgjlsfhk1 5y ago
        SF14's net was actually trained on lc0 data. Human data is way too weak for SF to learn from.
        • kohlerm 5y ago
          Yes and I think it's really generous of the LC0 to "donate" that data. On the other side LC0 uses SF during the training process since a while.
          • adgjlsfhk1 5y ago
            It makes sense. They're both open source programs working towards better understanding of chess. There are actually a few people who develop for both. They are obviously very different types of engines, but they are much closer to friendly rivals than enemies.
      • astrange 5y ago
        AlphaZero and co only trained via self-play because that was their research goal, and they looked unwinnable because they were the first to get this kind of neural net working, but it doesn't seem like it's the best option. AZ notably got completely tilted once it started losing, doesn't necessarily recognize strange positions you can't normally get into, and doesn't care about its win margin at all.