wrote in message
oups.com...
Although some of the hardware used for Deep Blue was remissioned for
other purposes, the software lives on. Since the project was completed,
no further investments have been made and it has not been enhanced. To
IBM, Deep Blue was simply a fun way to develop its concept of "deep
computing", that resulted in a much bigger PR payoff than expected.
There was no cheating. Garry could have drawn the final game, and
therefore the second match, but he took a calculated risk to try and
win it, that didn't work. Bravo for Garry ... he demonstrated that even
the largest amount of compute power that had ever been assembled for
chess could at best play even with the top human player.

I suppose that beyond the idea of human intervention [what was Benjamin
actually doing there, eg?] the second cheat foisted on the public was the
idea that the program beat the chess world champion.

How many moves did the program play out-of-the-book?

When did it begin to make out-of-book-moves?

Did it use end-tables? [and thus rejoin the book by using look-ups later in
the game]

[[look-up of material is not legal chess, btw, and cannot be compared with
legal chess]]

What was the technical role of the TD - ie, what were his terms of
reference?

If I had a massive database of all GM moves in the openings, then I could
confront any GM in the world with it, no? especially if it takes me no time,
but takes them an hour to negotiate my course to get to move 20. And this
would also consume GK's time and sap his energy.

GK played his weird anti-computer lines precisely to get around this
formidable problem - but his conceit was that this would advantage him more
than the program - which it did not. And so we have on a technical basis GK
playing daft chess, not his own game, or indeed chess-as-we-know it.

This is the second aspect of the cheating which IBM's apologists do not
address.

Phil Innes