| |
| |
All of the most important areas for expansion were mentioned elsewhere
in this document: improving the evaluation function, running the search
in parallel on multiple machines, and en passant captures.
|
| |
To conclude, let's review the project's goals and evaluate whether we've
been successful.
- The system must understand the basic rules of chess.
Success, expect for certain obscure positions.
- The system must know when the game is over for wins, losses,
and draws.
Success.
- The system must always choose moves that will maximize the
strength of its own position. This includes finding all
checkmate combinations within some pre-specified depth of
search.
Success. Testing of the alpha-beta search shows that all checkmate
positions within the search depth
- The system must choose each move in two minutes or less.
Success, for most board positions. The most complex and dynamic
positions sometimes take considerably longer.
|
|