As I indicated, the difference is that a computer examines
all possible moves, while (good) human players
only examine a few moves. They instinctly know what good
moves are, and don't waste time even considering bad moves.
How humans make this (pre)selection of moves to examine isn't
known, but it's good enough to beat computers who can examine
millions of positions per second.
Abigail | [reply] |
How do you think humans make this initial pre-selection? Through "knowing" which kinds of moves are generally "best", resulting in the highest valued position, from other positions. It is from memory of being in that position/arrangement before or a very similar position/arrangement. Or do you believe there is something more mystical occuring in the human mind other than rapid recognition of previous experience and logical reasoning during this process?
Is it this pruning that is the exhibition of intelligence? Or is it the fact the the heuristics are written in neurons instead of code?
I believe you have a misunderstanding regarding the heuristics used in modern chess playing programs. They do not examine all moves for the entire tree. Branches are pruned as early as they become valued such that they are determined not worth following, the system does not continue to evaluate "losing" positions and the subsequent moves.
Deep Blue employs a system called selective extensions to examine chessboard positions. Selective extensions allow the computer to more efficiently search deeply into critical board arrangements. Instead of attempting to conduct an exhaustive "brute force" search into every possible position, Deep Blue selectively chooses distinct paths to follow, eliminating irrelevant searches in the process. ( from this story)
It is a rare individual who can beat the best chess playing computers; one of the top 5 chess players in the world has tried repeatedly and has somewhere in the area of a 50% win rate against Deep Blue.
| [reply] |
How do you think humans make this initial pre-selection? Through "knowing" which kinds of moves are generally "best",
resulting in the highest valued position, from other positions. It is from memory of being in that position/arrangement before
or a very similar position/arrangement. Or do you believe there is something more mystical occuring in the human mind
other than rapid recognition of previous experience and logical reasoning during this process?
I've said repeatedly that how humans make this pre-selection isn't
well known. What do you expect from me, do a handful of ph.D's
in this afternoon and come up with an answer? But it *is* known
that humans do. Saying "oh, they just do it from memory" is
bypassing the problem. A computer can evaluate more positions
in a minute than human will ever encounter in a lifetime - just
letting it run for a few weeks would create a fantastic library; afterall
a computer can store far more in memory than a human can. But
if it was as simple as that, the strongest players in the world
would be computers, but they aren't.
A human,
OTOH, can certainly recognize *patterns* (in positions, tactics,
strategies, combinations, etc), a feat chess computers don't
really use.
I believe you have a misunderstanding regarding the heuristics used in modern chess playing programs. They do not
examine all moves for the entire tree. Branches are pruned as early as they become valued such that they are determined not
worth following, the system does not continue to evaluate "losing" positions and the subsequent moves.
I wrote previously, and you quoted that in one of your replies:
After the opening, they just explore all possible plies (a ply is a half-move) to a certain depth, evaluate the postions, and prune to select the
best moves so far.
| [reply] |