28 January 2017

Fischer and Deep Blue

After reopening this chess960 blog in last week's post, 'Everyone I Know Plays Chess960', I'd like to continue with another long quote. This one popped up a few weeks ago while I was researching Hans Berliner (1929-2017) on my main blog. It's from an April 1999 thread in rec.games.chess (RGC), Fischer interviews -- does he say anything relevant?
I've listened (unfortunately, at least in what I have heard) to a couple of the Fischer [radio] interviews. I was wondering if he has anything relevant to say? If so, I would appreciate knowing which interview, since I haven't heard it yet! Does he talk about chess? His legacy to the game? (Which these interviews, I would imagine, aren't helping him any.) Does he say anything positive?

A list of all(?) Fischer interviews can be found at Bobby Fischer Live Radio Interviews (tripod.com). Many of them exhibit Fischer at his worst and I didn't listen to any of them while preparing this current post. The RGC question received a long answer. It started,

Fifth interview concerns Botvinnik and Deep Blue:

• His rating adjusted for inflation would be "at least 2900" according to a Spanish magazine article in about 1992. He got his rating honestly and has "never even prearranged a draw". And concerning Kasparov's rating... he achieved it by prearranging games with Karpov and Russian Jew players...

• "Botvinnik, he spent years developing a chess computer. But he never made this chess computer public, he never played any games with it. I believe what he was really developing was a computerized system of prearranging games, of inventing games. You can prearrange games now at unbelievably fast speeds. It used to be a very tedious hard process to prearrange a realistic looking game. Now you can prearrange a game I think in a matter of minutes. They can prearrange a whole tournament I think in a few hours or days now with this hightech computer technology."

In response to a question, 'How do you look at Deep Blue then?', Fischer replied,

Deep Blue... it plays very well. I think that match that Kasparov played Deep Blue was a genuine match. What did everybody say about Kasparov's play? Ask anybody about his play. Read all the stories. Everybody said Kasparov was unrecognizable. Why was he unrecognizable? Because it was a real match. His moves were not prearranged. That's why he was unrecognizable. The Kasparov that you think you know, you don't know. The Kasparov you think you know is the Kasparov who's played all his prearranged moves in these prearranged games.

'If you were there to play with that computer would you do it?'

I would play Fischerandom chess. If there was a good offer, I would play Fischerandom chess, sure.

'Against Deep Blue?' • 'Yeah, yeah.' • 'Really. And you're not afraid about that?' • 'No, I'm not afraid, no. Why should I be afraid? The worst that could happen is you lose.' • 'But you can play against the machine. You know, quicker in response than a human being.'

Now here's the bit that's especially relevant to chess960 (aka Fischerandom, aka FischerRandom).

Yeah, but I think with Fischerandom chess the strategy is much more complex... The great strength, one of the great strengths, I won't say the only great strength but one of the very great strengths of the computers are to have an enormous opening library. And they can get a really good opening generally speaking. And the opening is the hardest part of the game. But they can get a good opening because they're drawing on millions and billions of hours of human manpower that has developed these openings over a couple of centuries, you see. They gain the benefit of intensive study by millions of chessplayers all over the world for 200 years. Now you take that all away from the computer, put the computer on its own in the openings I think the computer will probably just... The computer's great strength is calculation but there's not really very much to calculate the first few moves of the game, it's a very high level of strategy, you know.

And this... Maybe the computer's up to it, but I'm not sure it's up to it at all yet. This is really artificial intelligence, the early part of the game. Once you get out of the opening and the pieces start to make some kind of contact that is, threats and defence, then the game starts to play itself and then the computer's tremendous... It can calculate unbelievably well but in those early moves of the game where it's really high-level, intelligence is involved where there's no immediate threat, so you have to combine [what] you perceive maybe threats in the distant future with your early positional moves. You're combining tactics and positional chess. It really takes a lot of thinking and a lot of intelligence, I believe, I don't think it's so cut and dried as the old chess.

I believe if I were to play Deep Blue with Fischerandom chess, I'd get very good openings. And once I get a good opening, I'm pretty tough. The computer will find that out. On the other hand, on the other hand, Pablo, computers are getting better every day and they're getting more intelligent and they're getting programmed so I don't guarantee anything but I believe I could take it. I believe I could take Deep Blue based on what I saw of its play in its last match with Kasparov. Maybe they've improved it a lot, I don't know. I can take it in Fischerandom chess if it plays at the level it played against Kasparov."

The author of the RGC post summarized, 'He is absolutely finished with the old chess. "I don't play chess anymore... I'm through with chess." [...] "Prearrangement is just totally dominating chess now."'

Then the author made the connection with Hans Berliner: 'The comments on Botvinnik explain everything for Berliner. If he'd just waited until now, he wouldn't have had to publish in ICCA Journal.' This remark deserves an explanation, but goes beyond the objective of this chess960 post.

One last point in the RGC post is worth highlighting: 'The interviewer was **brilliant** to insert the comment about Deep Blue right at that point. Even though he sounds rather sleepy. But I suppose even though many people would like to see a Fischer - Deep Blue Fischerandom match, it will be like Kasparov - Shirov, everyone wants to see it but circumstances prevent it from happening.'

This met with an objection in another RGC comment: 'Here's the rub: Deep Blue's chess program was hardcoded into the chips -- there is absolutely no possibility of it playing Fischer at FischerRandom! Bobby undoubtedly knew this when making the above comments. So it was more hot air -- enough to fill a balloon.'

That last comment also goes beyond the objective of today's chess960 post. Maybe I'll continue it on my main blog.

No comments: