If I had a unicorn, DB would lose, though I’d have to sacrifice my unicorn’s kidney to escape a stalemate.
Chaaarrrliieeeee…
I can’t find a link for this, but I read a long article that argued (well) that contrary to common belief, Deep Blue’s victory over Kasparov doesn’t mean that computers can now outplay humans, for three main reasons:
-
The program was modified between games. If you don’t believe that this is significant, ask yourself why was it necessary to do this?
-
In the past, the difficulty hasn’t been in developing a system that can beat a grandmaster. The difficulty has been in making a system that can continue to do so – experts eventually notice patterns in the system’s play: emergent behaviours manifested from the particular way the algorithm is traversing the search space.
Deep Blue was disassembled soon after beating Kasparov. -
Commentators say Kasparov played badly, and should’ve at least drawn the match but for a couple of uncharacteristic mistakes.
– I see that there has been a documentary film on this subject –
It should be noted that the match where Deep Blue beat Kasparov was sponsored by IBM, who therefore set the schedule. And the schedule they set was much more taxing than usual, which of course favors the contestant which doesn’t get tired.
Coffee now spewed on monitor. Thanks.
But what if the “corn flake” shapeshifter was able to fight itself?
But yes, I remember Archon well. I only was able to track it back down again, because of said “corn flake” piece.
However, Im still not clear on 'charmed squares being proof against magic". ((Meeko watches as the thread gets hijacked. ))
As I understand, there is effectively some randomness to DB’s moves, in the sense that the programming isn’t “Find the best move” but “Find the best move in what, to humans, is a reasonable amount of time”, i.e. at a pace of ~60 moves in 60 minutes, about the minimum speed for a tournament game. Deep Blue’s speed is sometimes given as 11.38 Gigaflops, but this isn’t a precise number. The room’s temperature might have a small but significant effect on the electronics, to where DB gets 11,380,193,824 positions in a second or 11,380,978,102 positions in a second. Both are just best guesses of the available outcomes (which in chess could be 10[sup]50[/sup] and upwards), but the second is slightly “deeper” and (in theory at least) leads to better play.
If the computer had infinite time or infinite speed with which to calculate its moves, it would play perfectly. I don’t know if white’s first move has ever been proven to be an advantage, but I figure DB playing itself perfectly would either get a win for white every time (if it is) or a draw every time (if it isn’t).
If there is no basis for assuming that playing white gives you an advantage, or that Deep Blue plays better as white, then there is no reason for assuming that white would always win. Chess is a completely deterministic game (given infinite time and memory, that is), and there are many completely deterministic games where only the second player is guaranteed a win.
Fine, if it can be proven that white making the first move is a disadvantage, then perfect play should lead to black always winning. Whatever.
While this is true in theory, in actual practice, white wins significantly more often than black in high-level chess, making it plausible to assume that white has an advantage. Further, while chess doesn’t have any true stalling moves, it does have some options which come very close to stalling, and in any game where it’s possible to stall, if anyone can force a win, it’s the first player.
Where the heck is glee, anyway?