Leonardo is currently at about 1900elo in Lichess https://lichess.org/@/Leonardo-Bot (It may not be online, since I don't have the means to host it 24/7
- Opening Book
- Iterative Deepening
- NegaMax
- Transposition Table
- Quiescence Search
- Killer Move Heuristic
- Butterfly History Heuristic
- NMP (Null Move Pruning)
- SSE (Static Exchange Evaluation)
- MVV LVA (Most Valuable Victim, Least Valuable Attacker)
- Selfmade NNUE (Efficiently Updatable Neural Network)
- Principal Variation Search
- LMR (Late Move Reduction)
- etc
Another Project of mine is a neural network (https://github.com/EliasKramer/NeuroFox), which I incorperated into this project. The most important features of a NNUE (e.g. partial forward feeding) were implemented. The difference to other NNUEs is that it only takes pawn structure as input. This makes it a lot faster than a common NNUE, but it doesnt boost the elo that much.
This project has probably a serious bug, because a Bot with these features should be in the range of 2700 elo