Next Steps

[tildedave]

With evaluation focused on material it is hard for the search to get out of the earlier search phases. More implementation here will help the search prune what a good or bad position is more effectively which means it should be able to search deeper.

Search

• Futility pruning and SEE in quiescent search

Evaluation

• Passed pawn detection, give plusses to those
• Give minuses to pawns that can't advance
• Bishops and knights as "good"
• Prioritize pieces being developed during the opening/middlegame

Move Generation

• Explicitly only generate captures (special logic)
• Switch move generation to not create new slices (will need some way to do this with "priority" for move generation)

[tildedave]

Will re-evaluate everything. Did some of these already, will do others later.