I've built a standard chess AI and trying to add a knowledge base. A rule based system beyond negascout and eval. It has all the standard parts, and now I want to dabble in novel directions. Tried chessprogramming.com but not any pseudo code for using a knowledge base, just heuristics. Opening / endgame table spaces are too scripted but inspiring. So are piece-square tables. Both seem too rigid and force a curtain game type.
Komodo is said to have this.
Best thing I can think of is a series of "if pattern then else action" similar to Alicebot Chat AI. Any ideas of how to implement a knowledge base that works inside an eval or negascout. Something that adds to selectivity or beyond the normal eval heuristics.