Re:Here's a real "what's the point" question:
on
Men vs. Machines
·
· Score: 1
We devote about 50% of our resources to implement teaching and training features in Fritz -- essentially to make the program weaker, but in a realistic, human fashion. Example: the "Sparring" mode in which the program tries to find moves that allow the opponent to win something with a clever combination (the regular playing mode never allows that). Afterwards it shows you all the places where you missed the tactics it had set up for you and quizzes you on the correct move to play.
With Fritz7 the main intention is to get people to play people on the Fritz server. So one full half of the program is devoted to logging into the chess club and playing against people all over the world or following broadcasts of important events.
It is only when you go to the other half and tell the engine to play full strength with no holds barred that you get wopped. Unless your first name is Garry, Vlady or Vishy. Then you just get a very good fight.
First some clarification: I am a very poor otb (=over the board) chess player, my highest rating ever was about 2000 as a 19-year-old. I speak exactly two languages, English (my first) and German, but both of them proficiently enough to write and publish in. I am a computer chess expert but not a programmer. Garry has been a close friend since 1985.
In my news blurbs on the main ChessBase page I try to make every item at least theoretically amusing or interesting to visitors, even if they are not chess players. Quite a daunting task, you must admit.
Initially I wrote "Michael Jordan tried it with baseball - it, er, did not work out." An Italian visitor drew attention to the "typo". So I changed it to "it, well, didn't work out". But then I remembered this incredible turn of phrase used by American kids these days. I had recently discussed the exact meaning and usage with an American colleague, Rudy Chelminsky of Wired et al., who abhors it's current prevalence. Rudy also introduced me to the really crass form: he was like "Never ever use it, Fred!"
So I changed it again. And apparently it worked. Look at the number of posting here dedicated to one word. That's good Fleet Steet technique - get them to look twice. And I learnt a new technical term: valley girl speak. I wondered why the Fullbright students here in Germany never use it. They're like all from New England!
Currently the world's number 15 human, Ilya Smirin, is playing against four of the world's top programs (info). He is well acquainted with the style of computer play, understands the strengths and weaknesses of the machines and prepared carefully for this match. In most of the games he has outplayed the programs, but is only one point ahead in the seven games played so far.
Tomorrow (Sunday) is the last game of the series. One has to be repeated after a very unusual incident: Deep Junior was winning but the Internet connection broke down and the computer could not process Smirin's move. So the operator offered him a draw. Smirin refused, saying he did not deserve to share the point. Instead he offered to resign. The Junior team refused because the program had not demonstrated the win. So they decided to repeat the game (info).
"Both Shannon and Turing spent quite a lot of time on chess algorithms. Shannon actually wrote the first chess program before a computer existed. He 'ran' the program using slips of paper and generated moves this way."
Actually it was Alan Turing, who wrote and operated the "paper machine" -- he played the role of a human CPU. A brief history of computer chess, including the game the poor man played, is available here.
"The chess programming breakdown already posted is pretty good. The key concept these days is brute force speed versus knowledge. 20 years ago most programmers thought you needed to make the thing somehow think like a human because the brute force method was so slow. Intel and Moore won. The "fast searchers" now dominate thanks to the minimax algorithm. It just looks at one line after another and counts the beans to rapidly prune. Programs differ to an extreme degree in the amount of knowledge they apply. (HIARCS, for example, is one of the few "slow" programs at the top. It applies a lot of knowledge and looks at maybe 1% of the number of positions the fast programs like Fritz and Junior check.)"
It's more like 20%. On a 666 MHz Pentium this is the speed (in thousands of positions per second) of the current top programs:
Note that Fritz7 has slowed down over the years. This is because it now has a lot of general chess knowledge built in. But you cannot measure knowledge by lack of speed. Fritz7 has more knowledge than Hiarcs 7, which is a number of years old. In fact it has more chess knowledge than any other top program available today.
"Those who think chess is solvable should speak only theoretically. The number of positions is one of those great "million times the number of stars times the grains of sand in the world" numbers."
Even the number of elementary particles in the universe (10^80) is a trivially small number, silly and insignificant compared to the number of possible games up to move 40 (10^112). But the number of unique positions that can occur on a chessboard is much smaller: 10^40 (just as the number of different words in a language is much smaller than the number of potential messages that can be generated from them). These can be theoretically solved using the Thompson back-solving method you mention. But storing the results would require the matter contained in millions of galaxies, so the game is unsolvable for all practical purposes. Just imagine what Greenpeace would say if they discovered we were dismantling millions of galaxies just to store chess!
The last part of your posting seems to have been cut off. Pity.
PS: For the others: I'm Frederic Friedel and part of the Fritz team.
Correct. If every particle in the universe was a super-computer and generated, evaluated and stored a billion positions per second, and all had started immediately after the Big Bang, and all were working with a perfect parallel algorithm, i.e. no computer duplicates the work of another, then the project manager today would have to admit that we still haven't scratched the surface of the problem. And that we're not going to solve it. Remember that in a mere 10^14 years all galaxies will have disintegrated and all stars burned out. After that the residual matter will also evaporate, so that after about 10^80 years only electrons and positrons will be left over. After 10^100 years even the super massive black holes will have evaporated. (After that things become hazy for me).
However: it is imaginable that we (Fritz 10 or Fritz 15) will find a way to force mate after 1.e4 in less time than this. This is analogous to a computer finding a forced win in a specific five-piece endgame position without knowing all 120,000,000 positions that can occur in that ending, and certainly without generating all possible five-piece games that can be theoretically played with that material on the board.
I have bad news for you. Chess will not be "solved" in the course of the current universe. Naturally we will not be able to generate and store all possible chess games, even up to move 40. That would be 10^112 games, which reduces the number of elementary particles in the universe to a silly, irrelevantly small number.
But of course we could take the other route: generate all possible positions and calculate all winning lines, as Ken Thompson did in the 80ies for three, four and five-piece endings. When you complete the retrograde analysis on these the computer plays them perfectly. We even use these endgames in the Fritz search, so that these days typically Fritz ends its won games by announcing mate in 20 to 70 moves. The first batch of six-piece endings have now been generated, but they are big. And the interesting ones (not yet calculated) are with pawns, which makes them even bigger. Pawns ruin one plane of symmetry and have the irritating habit of being promotable, so that you have to have all the piece endings that can derive from them available as well. Fritz can access a number of pawnless six-piece endings, but there is currently no practical use for installing these endgames.
How about the 32-piece endgame? Now this is seriously big, containing between 10^40 and 10^50 positions. Trust me, it cannot be handled by the computer farms you propose. Dr John Nunn (mathematician, chess Grandmaster) tells me that just storing the 32-piece endgame database using optimum storage techniques (smallest theoretical energy levels for each bit of information) would require the matter stored in a few million galaxies. Maybe it could be done, but John is worried what Greenpeace would say about a project that involves dismantling galaxies in order to store chess.
They'd probably chain themselves to the rocket ships.
Re:Fritz's last thoughts?
on
Mir Deathwatch
·
· Score: 1
Actually I did a final interview with Fritz, just a few hour before the Mir was deorbited. Poignant, curious, interesting. I will be publishing it sometime next week at http://www.chessbase.com
Slashdot Mirror
We devote about 50% of our resources to implement teaching and training features in Fritz -- essentially to make the program weaker, but in a realistic, human fashion. Example: the "Sparring" mode in which the program tries to find moves that allow the opponent to win something with a clever combination (the regular playing mode never allows that). Afterwards it shows you all the places where you missed the tactics it had set up for you and quizzes you on the correct move to play.
With Fritz7 the main intention is to get people to play people on the Fritz server. So one full half of the program is devoted to logging into the chess club and playing against people all over the world or following broadcasts of important events.
It is only when you go to the other half and tell the engine to play full strength with no holds barred that you get wopped. Unless your first name is Garry, Vlady or Vishy. Then you just get a very good fight.
Ahh, ONE person who understood.
First some clarification: I am a very poor otb (=over the board) chess player, my highest rating ever was about 2000 as a 19-year-old. I speak exactly two languages, English (my first) and German, but both of them proficiently enough to write and publish in. I am a computer chess expert but not a programmer. Garry has been a close friend since 1985.
In my news blurbs on the main ChessBase page I try to make every item at least theoretically amusing or interesting to visitors, even if they are not chess players. Quite a daunting task, you must admit.
Initially I wrote "Michael Jordan tried it with baseball - it, er, did not work out." An Italian visitor drew attention to the "typo". So I changed it to "it, well, didn't work out". But then I remembered this incredible turn of phrase used by American kids these days. I had recently discussed the exact meaning and usage with an American colleague, Rudy Chelminsky of Wired et al., who abhors it's current prevalence. Rudy also introduced me to the really crass form: he was like "Never ever use it, Fred!"
So I changed it again. And apparently it worked. Look at the number of posting here dedicated to one word. That's good Fleet Steet technique - get them to look twice. And I learnt a new technical term: valley girl speak. I wondered why the Fullbright students here in Germany never use it. They're like all from New England!
Currently the world's number 15 human, Ilya Smirin, is playing against four of the world's top programs (info). He is well acquainted with the style of computer play, understands the strengths and weaknesses of the machines and prepared carefully for this match. In most of the games he has outplayed the programs, but is only one point ahead in the seven games played so far.
Tomorrow (Sunday) is the last game of the series. One has to be repeated after a very unusual incident: Deep Junior was winning but the Internet connection broke down and the computer could not process Smirin's move. So the operator offered him a draw. Smirin refused, saying he did not deserve to share the point. Instead he offered to resign. The Junior team refused because the program had not demonstrated the win. So they decided to repeat the game (info).
Hello Michael,
I have some comments. You write:
"Both Shannon and Turing spent quite a lot of time on chess algorithms. Shannon actually wrote the first chess program before a computer existed. He 'ran' the program using slips of paper and generated moves this way."
Actually it was Alan Turing, who wrote and operated the "paper machine" -- he played the role of a human CPU. A brief history of computer chess, including the game the poor man played, is available here.
"The chess programming breakdown already posted is pretty good. The key concept these days is brute force speed versus knowledge. 20 years ago most programmers thought you needed to make the thing somehow think like a human because the brute force method was so slow. Intel and Moore won. The "fast searchers" now dominate thanks to the minimax algorithm. It just looks at one line after another and counts the beans to rapidly prune. Programs differ to an extreme degree in the amount of knowledge they apply. (HIARCS, for example, is one of the few "slow" programs at the top. It applies a lot of knowledge and looks at maybe 1% of the number of positions the fast programs like Fritz and Junior check.)"
It's more like 20%. On a 666 MHz Pentium this is the speed (in thousands of positions per second) of the current top programs:
Fritz7: 300 kN/s
Fritz6: 450 kN/s
Fritz5: 520 kN/s
Shredder: 160 kN/s
Junior7: 250-435 kN/s
Tiger: 135 kN/s
Hiarcs7: 65 kN/s
Note that Fritz7 has slowed down over the years. This is because it now has a lot of general chess knowledge built in. But you cannot measure knowledge by lack of speed. Fritz7 has more knowledge than Hiarcs 7, which is a number of years old. In fact it has more chess knowledge than any other top program available today.
"Those who think chess is solvable should speak only theoretically. The number of positions is one of those great "million times the number of stars times the grains of sand in the world" numbers."
Even the number of elementary particles in the universe (10^80) is a trivially small number, silly and insignificant compared to the number of possible games up to move 40 (10^112). But the number of unique positions that can occur on a chessboard is much smaller: 10^40 (just as the number of different words in a language is much smaller than the number of potential messages that can be generated from them). These can be theoretically solved using the Thompson back-solving method you mention. But storing the results would require the matter contained in millions of galaxies, so the game is unsolvable for all practical purposes. Just imagine what Greenpeace would say if they discovered we were dismantling millions of galaxies just to store chess!
The last part of your posting seems to have been cut off. Pity.
PS: For the others: I'm Frederic Friedel and part of the Fritz team.
Yes, a number of pawn-less six-piece endings have been compiled. See my other postings below.
Correct. If every particle in the universe was a super-computer and generated, evaluated and stored a billion positions per second, and all had started immediately after the Big Bang, and all were working with a perfect parallel algorithm, i.e. no computer duplicates the work of another, then the project manager today would have to admit that we still haven't scratched the surface of the problem. And that we're not going to solve it. Remember that in a mere 10^14 years all galaxies will have disintegrated and all stars burned out. After that the residual matter will also evaporate, so that after about 10^80 years only electrons and positrons will be left over. After 10^100 years even the super massive black holes will have evaporated. (After that things become hazy for me).
However: it is imaginable that we (Fritz 10 or Fritz 15) will find a way to force mate after 1.e4 in less time than this. This is analogous to a computer finding a forced win in a specific five-piece endgame position without knowing all 120,000,000 positions that can occur in that ending, and certainly without generating all possible five-piece games that can be theoretically played with that material on the board.
I have bad news for you. Chess will not be "solved" in the course of the current universe. Naturally we will not be able to generate and store all possible chess games, even up to move 40. That would be 10^112 games, which reduces the number of elementary particles in the universe to a silly, irrelevantly small number.
But of course we could take the other route: generate all possible positions and calculate all winning lines, as Ken Thompson did in the 80ies for three, four and five-piece endings. When you complete the retrograde analysis on these the computer plays them perfectly. We even use these endgames in the Fritz search, so that these days typically Fritz ends its won games by announcing mate in 20 to 70 moves. The first batch of six-piece endings have now been generated, but they are big. And the interesting ones (not yet calculated) are with pawns, which makes them even bigger. Pawns ruin one plane of symmetry and have the irritating habit of being promotable, so that you have to have all the piece endings that can derive from them available as well. Fritz can access a number of pawnless six-piece endings, but there is currently no practical use for installing these endgames.
How about the 32-piece endgame? Now this is seriously big, containing between 10^40 and 10^50 positions. Trust me, it cannot be handled by the computer farms you propose. Dr John Nunn (mathematician, chess Grandmaster) tells me that just storing the 32-piece endgame database using optimum storage techniques (smallest theoretical energy levels for each bit of information) would require the matter stored in a few million galaxies. Maybe it could be done, but John is worried what Greenpeace would say about a project that involves dismantling galaxies in order to store chess.
They'd probably chain themselves to the rocket ships.
Actually I did a final interview with Fritz, just a few hour before the Mir was deorbited. Poignant, curious, interesting. I will be publishing it sometime next week at http://www.chessbase.com
Hello, I'm Vlada .
Send me mail, Joel. Maybe we can get something cool together. [frederic@chessbase.com]