-- Via DLG Pro v0.995
Daniel Marsalone Daniel Marsalone Daniel Marsalone
Internet: daniel_m...@agwbbs.new-orleans.LA.US
UUCP: rex!agwbbs!daniel_marsalone
The fact that there are fewer pieces doesn't allow the computer to search
many more ply ahead. The search tree grows extremely fast. Perhaps you've
watched some chess programs like Sargon "think." If you look at the ply, you
notice that the ply increases quickly at first and then gets slower and slower.
The effect of fewer pieces is not nearly enough to counteract the expanding
tree at higher plies, so the the same amount of thinking time may only yield
an extra 2 or 3 ply.
Another thing to keep in mind.... Depending on how the chess algorithm
works, it is easy for the program to miss things in the endgame that are
obvious to humans. The easiest way to illustrate this is to consider a
situation where you have a couple pawns at their starting positions and
so does the computer. There are four pawns on the board and each is 5 or
6 moves away from queening. Even if there are just two such pawns on the
board (one for each player), the computer may have to look 12 ply ahead to
see who queens first and what happens immediately afterwards. I'm
exaggerating things a bit here since I believe current chess algorithms are
a bit smarter than that, but the point is that things that may be obvious to
you may not be obvious to the computer and there is not a very significant
increase in search depth due to the lower number of pieces on the board to
compensate for the computer's stupidity in the endgame.
-Andy
P.S. My connection to my MIT account is down so I'm writing from one of
my other accounts with the wimpy newsreader. It's so easy to get spoiled
with Xwindows....
P.P.S. It's hard to get used to the fact that we can now talk about
chess on rec.games.chess. What a concept!