Chess. (Part II)

So let's talk about strategy.

What is strategy exactly? Well I have always thought about it as a set of imposed rules that are beneficial towards some goal. For example when someone plays poker, they may have a set of "rules" they tell themselves like "never bet your life savings on a bad hand" which are not explicitly part of the rules of the game.

Now as I have said many times, computers are idiots, you need to spell out EVERYTHING to them. So how can we teach them a proper strategy?

Well the answer is to take into account EVERY contingency.

Now to apply this to chess. Take a look at this board:

If you look at it for a while, you may be able to see that, as white, you can checkmate the opponent in one move (Bg4-f3). However for a computer to get to this point, it would need to have a very complicated set of instructions. First it might try a bunch of legal moves for every piece, so maybe it would move the knight to b6, as this would provide a check, however then the king could escape to b7 and you would no longer be able to checkmate so easily.

So the key is teaching the computer some kind of point system to evaluate the moves. One of the ways this is done is to think of various pieces as having point values:

pawn - 1
bishop and knight - 3
rooks - 5
queen - 9
king - over 9000

So while the computer is checking possible move permutations, it might get black to try and capture the knight if you played b6, but then as white you would be able to capture the king and thus get 9000 points. Also as it tests various lines, it would stumble upon checkmate in this way since it would find an unavoidable line that results in the king being captured.