Chess Notes

For the Purpose of Writing

An Alogrithim that Can Be

Coded In A Computer Language

December 27/2012

Opening Procedures

Middle Game Procedures

End Game Procedures

It is useful to have the enemy King in check every move.

CheckMate Search Procedures

Overall Game Procedures

Use a pawn or lower value piece as point with lots of high level backup.

Moves that also place the enemy in check may be beneficial.

Attacking two pieces with one piece MBB.

Having the rooks linked MBB.

Make sure your pieces have freedom to move in the future if possible. Blocking the opponents freedom to move or choosing a move which leaves remaining a block that is already there.

Squares in the center are more valuable with a decreasing value outwards.

Defensive and offensive positional qualities will have to be observed, such as a line of pawns, and that they need to be taken out at the base, changing the value of that particular square.

Pieces should go out as teams of two or more.

Controlling as many squares as possible, even ones not in use. The squares will have to be rated for value each turn.

Other Notes

This algorithim will not use memorized or rote procedures because that is what I want. Every move is to be reasoned out, using a imitation of my thinking procedures for chess.

I wish for this to be fast. I will do this by not searching every move combination, which is time-consuming but by searching rationally for the best move.

I will limit the algorithim to searching only 3 deep (?) for moves. This will limit the amount of moves the AI has to rate.

Each factor in the decision making process will be given a statisical rate of value. Some of these values might vary because of links to other factors. Likely, the more factors and the better I can rate them, that will improve the performance of the AI.

The goal is not necessarily to make the best chess engine but one that most accurately reflects my mind. Although with a mistake rejection system. And see how it does.