I think we have slightly different views in this one. I hold the view that chess power in computers is due BOTH to brute force AND good algorithms (the evaluation function).

As for whether IBM has better programers than Junior (a software company), I did not claim that. See, Deep Blue was a specially inbuilt hardware. Supercomputer on a chip if you will. Its strength came mainly from the much higher brute calculation speed. I will concede that if Junior was running at the same speeds, it may very well be stronger than DB. As for Hydra, it is a hardware like DB, but with newer tech, the 64bit chip etc. It runs at similar speeds as DB did 9 years ago. Is it better than DB of 9 years ago? Maybe (I dont think so, but I will not go into details here; I intend to post comments on the pages related to the match with Adams) Does it have better programmers than DB did 9 years ago? I dont think so. Their team is rather small and it doesnt have the vast resources IBM could afford.

Could IBM build an updated version of DB today, using the newer chips and refining their past work on positional evaluations, running at 100 times the speed DB run in 1997 (100mil/sec in closed positions and 200mil/sec in open positions)? Yes! Would this DB beat Hydra, Kasparov, Junior etc? I believe it would. It would ``see`` a lot deeper than them and it would analyse everything within its event horizon, without the need for pruning. I think a newer DB crunching 10-20 billion/sec would run over Hydra, Kasparov, Kramnik or Anand.

I'll explain. Because of the nature of chess, examining what we call the "complexity tree" of a position is something that get's pretty ugly after just a few moves. There are about 20 possible moves for the first player to move and then 400, 8902, 197742, 4897256, 120921506, 3284294545 ... As you can see, after just 3 moves a player, the number of possible moves is 3 bilion! (http://mathworld.wolfram.com/Chess....). Ofcourse with less pieces on the board the numbers are smaller but the argument holds for the opening, middle-game and even for some end games situations. There is a point in which even if you make the computer 1000 times faster, it will not see even one more move ahead of a slower version, because of that complexity.

Computers, due to their "stupidness" can't evaluate a position as a human would, and they can't rely on instincts for the best possible move, or for guessing their opponet's plan. They also can't process the entire game complexity tree (see 12 pages of above argument for solving chess). Therefore, they must rely on what we called the "evaluating function" - the "core" algorithms - to evaluate the position.

<As for Hydra, ... , Is it better than DB of 9 years ago? Maybe (I dont think so, ...>

I am the last person in the world to judge the strength of a computer, since I lose to the pathetic Chessmaster 4000 (and even not in it's highest level!)... But here is from wikipedia:

(http://en.wikipedia.org/wiki/Chess)
"The chess machine Hydra is the intellectual descendant of Deep Blue; and appears to be somewhat stronger than Deep Blue was. Certainly it is very much comparable in terms of positions analysed per second."

<Their team is rather small and it doesnt have the vast resources IBM could afford>

I agree completely. I think that every human that goes on to take a strong computer in chess needs to see a psychologist.. If he want a headache I can arrange it for him more easily. :)

In the end, no matter who/what you play, playing the game is practice. And this practice will make you stronger if you understand the true nature of that practice (ie recognize that humans aren't computers).

The things I like most about being able to play computers: they're always there when you want to play, they don't mind restarting or taking back moves, they'll play on from any position, and they don't laugh when you blunder.

that remark about Morphy is interesting. In retrospect many combinations turn out to be flawed, either losing by force, or leading to a draw instead of other quieter moves that might have lead to a win. That is true of many games of Alekhine, Tal, Kasparov etc. Combinations are hard to calculate over long depths. If your instinct is wrong then you have a flawed masterpiece. What you said about open games is also very true. Playing an open game against a powerful computer is asking for trouble. Not that computers are gods, but any small error would be spotted and if your judgement was flawed, then the computer will repel your visually compelling, but analytically wrong attacks, keep the extra material and grind you to defeat.

Maybe that is why Kasparov played the types of openings he did in 1997. In game 3 he went for a fortress type position (like Adams in his 2nd game against Hydra) and had to go for a draw. I will post in that link (game 3) my thoughts on that one.

If you look at his game 1, his pieces seldom passed his third rank. Maybe that is the only way to beat the supercomputers: Petrosian-style refined position play. Look at the bad result Adams had in normal, open games. Of course, Adams is not as strong as Kasparov, but the score was attrocious. Playing open games (against a supercomputer) is an invitation to disaster. I doubt (if a rematch were made and Kasparov kept his promise to play 1.e4 and the Sicilian) he would have done any better. That would have played into the strengths of the computer and lead to tactical play where Kasparov would have been squashed by DB. (just remeber game6: if the slightest tactical chance is given, the supercomputer runs over you. That may have been an opening blunder, but DB made it look easy in its squashing punishment. Many openings with 1.e4 and many Sicilian variations contain such gems, making DB or Hydra a monster to play against in such openings.)

Maybe the best results will be between positional players using closed openings.

<In the end, no matter who/what you play, playing the game is practice. And this practice will make you stronger if you understand the true nature of that practice> I agree to some extent - what's the point in practicing towards something you can't win? If your point is that computers should be used for practice perpuse, I agree (again - to some extent) but Human/Machines clashes seems to favour our stupid machine friend..

<Surely you can't actually believe that <Machines should be banned from the game (with humans)!>?> Well, I am not a big fan of those Human/Machines matches, I think it's pretty silly, like throwing your head at a wall thinking you are practicing Kung-fu.. Ofcourse you can't break the wall so it's the best training! :)

But I'd like to restrict and say that computers have an important role in the shaping of science, and if chess is regarded as a science afair then I am most pleased with their progress...

It can be a freeing thing to recognize your own limitations. For a lot of people, it prevents them from throwing their lives away needlessly, to learn a game.

<It can be a freeing thing to recognize your own limitations. > It's familiar in some way :)

"Chess is 99 percent tactics."

A computer can perform these tactical calculations far better than a human can. That's just the reality of the situation. Chess had a nice run, from Greco to Kramnik, but in a few years, it will be game over, when a computer program you can buy for $19.95 can blow the doors off any human out there.

<There are far better ways to spend your life. like what? Pinball? Poker?>

Maybe with the Day of the Tenticales.
- "I feel like I could... <gluk gluk> I could! <gluk gluk gluk> Take on the world!" - "But now, I know that I must go... back to the mansion!"

This is nonsense, actually. If it were really true, computer programs would have been far better than the strongest humans a long time ago. It's hardly even true on average club player level, and definitely not even close on super-GM level.

The reason humans can still put up such good resistance is because there is just the right balance between strategy and tactics. And conversely, if tactics were not important we would beat the machines easily.

Quotes like "chess is tactics", "chess is 99% tactics" and so on are often meant, I believe, to mean that you can't win a game without some kind of tactics. That's, for example, the essence of Mig Greengard's <It was clear that no matter how much Chernev tried to break down every move, at the end of the day some guy made a mistake, the other guy punished it, and the only way that happens is tactically. That’s what chess is. Chess is tactics.> http://www.chesscafe.com/text/skitt...

< Let's take the general case - you have two objects, with masses m1 and m2. Each "feels" the force F1=G*m1*m2/r^2. The first object's acceleration is F1/m1, the second object's acceleration is F1/m2. So far, it's obvious. Now, let's multiply m1 by 10. F2, the new force, will be 10*F1. The first object's acceleration will be, therefore, 10*F1/10*m1 = F1/m1, it won't change. The second object will have it's acceleration increased by a factor of 10, of course.>

I don't think you're expressing yourself clearly. Your original expression is for the force of gravitational attraction between the two masses, in the absence of any other influence. Then you ring in the Earth? And previously you were discussing air resistance? I think you are only going to confuse your audience.

Maybe a simpler way of demonstrating your point would be to:

1. Do the old 2-line proof that the force of gravitational attraction between any mass and the Earth gives rise to the same acceleration in the same location, INDEPENDENT of mass. Newton forever.

strategy is the overall plan, i guess a VERY VERY basic strategy would be "I want to arrange as many pieces as I can on the e file".

tactics are the execution of that plan. "I will move my pond forward two spaces, thus opening more free spaces closer to me on the e file for my other pieces"

Tal made the comment that in his matches with Botvinnik - Tal was calculating, calculating, calculating. Botvinnik was not so much of course - be he was thinking schematically - "in this position I must castle queenside." What is interesting is, Tal said, he and Botvinnik was arriving at the same conclusion, more often than not, through different methods.

So, if you take Tal's method, and multiply it many times over - if doubling rooks on the e file is the best approach, the computer is going to see that too.