Practical move selection in a Stockfish-based engine (project idea)

[peio53]

Hello,

I’m working on a personal chess engine project based on Stockfish, but with a different goal than pure objective strength.

The idea is to select, among objectively sound moves, those that are most difficult to face for a human opponent (maintaining tension, creating non-forcing threats, limiting natural replies, etc.).

The project does not aim to modify the search or NNUE, but to introduce an additional decision layer for move selection.

I’ve written a clear concept/spec document describing the philosophy and intended behavior, and the project is planned as open-source.

At this stage, I’m mainly interested in discussing the idea with people familiar with chess engines, to see whether the approach makes sense technically and conceptually.

If the concept sounds interesting, I’d be happy to share the document and exchange ideas.

Best regards.