It currently process around 30k nps and uses a combination of tactical and strategic evaluation. It's current playing strength is around 1650-1750 (my guess) depending on the type of game. I have hit a few issues and was googling for ideas when I found this site.
- 0) How do I ensure that the program is playing better after each change I make?
1) The program can get stuck in some positions. It's ahead but it needs to sacrifice material etc to break through. Obviously a deeper search would help but are there any techniques for spotting that a change in plan is required. At the moment it just moves it's king back and forwards hoping for divine intervention.
2) 30k nps is ok but it's not allowing me to look deeply. I'm currently looking into different ways of extending the search on exchanges. I'm also trying to understand the concept of null moves to improve alpha beta pruning but it's a bit uphill at the moment:) I also wondered about widening (?) the alpha beta cut off point so it just ignores anything that's even remotely close to the current best score. Ideas?
3) I'm also looking for a new name for the program - I guess "fun chess" has been taken.
Mike Robinson