< Earlier Kibitzing · PAGE 54 OF 54 ·
|Jun-26-05|| ||Dionyseus: <acirce> How about you post a line and contribute to this beautiful rook pawn endgame instead of criticizing me? Btw, I believe it can be argued that me + my program can beat humans in analysing rook endings. Let's not forget, even you have praised some of my endgame analysis recently.|
<Honza Cervenka> Your 36.h3+ is very nice, gave me lots of problems, but I think I have finally found a refutation. Instead of 37...Rd3, I suggest black starts rolling his pawns down to victory with h6!
So 33.Re6 Kg4 34.Kg2 Rc3 35.Rxd6 f4 36.h3+ Rxh3 37.Rd7 and now h6. What can white do to stop black from advancing his G and H pawns?
If white plays 38.Rh7, black simply continues g5 39.kf1 kf3 40.kg1 g4 41.Rb7 Ke4 here it is obvious black can no longer defend his D pawn. 42.Re7+ Kxd5 43.Kg2 h5 44.Ra7 Rd3 45.Rh7 ke4 46.Rc7 h4 and there's no need to go further, it's clear black wins.
|Jun-26-05|| ||Honza Cervenka: <Dionyseus> After 33. Re6 Kg4 34. Kg2 Rc3 35. Rxd6 f4 36. h3+ Rxh3 37. Rd7 h6 38. Rh7 g5 I would have prefered 39. d6 Rd3 40. d7. Then my Fritz has still troubles to beat me: 40...f3+ 41. Kh2 h5 42. Re7 h4 43. Re4+ Kf5 44. Re3 Rxd7 45. Rxf3+ Ke4 46. Rf8 etc. looks still defendable.|
|Jun-27-05|| ||Dionyseus: <Honza Cervenka> Looks like you're right. After 39.d6, I can't find a win for black.|
|Jun-27-05|| ||lostemperor: Statement: If given unlimited time to think any chessplayer can beat a computer.|
I came to think about this through a friend of mine. I knew he had a system he always beat his computer chessmaster 9000 with it. a strange 1.e3 2. d3 3.h3 4.a3 system what will be very bad against humans but it worked against his strong computer! He can also do this system with the black colors. Of course against a computer like Hydra this system will not work but it's the thought behind it, that counts.
He did say 1.e4 always loses. I did not understand what he meant. Later I did. He meant that if he plays the normal 1.e4, instead his 1. e3 system, he always loses. One day (recently) he came to me and said "I finally won with 1.e4! It costed me a lot of take backs but I did it!"
About take backs: I think a top-player may only need 1 or 2 take backs possibility in each game to draw (or beat) the computer. Average clubplayers, like me, may need 10 to 20 take back moves to draw (or perhaps beat) the computer.
So if given an unlimited time to think almost anybody can beat the computer. In fact this has now recently been proven in the freestyle tournament where one can consult another computer but it would work without consulting another computer either, and also by correspondence games where GM. Nickel beat Hydra.
|Jun-27-05|| ||catfriend: Nickel used computers to do it...
And don't forget to give this umlimited time to the computer as well, making it almost perfect!
|Jun-28-05|| ||lostemperor: Sure Nickel used computers. He would not stand a chance in an OTB game (but others may). But even without computers if given enough time any chessplayer can draw Hydra. I believe even beat it. To give the computer unlimited time won't change matter much since it is already so fast it won't get much more "understanding" of the position as man. I'm not exactly sure but increasing the speed tenfold would only make it think one ply deeper at most (and decreasing).|
|Jun-28-05|| ||supertimchan: No.... Even if you give a club enough time, without computer, they can't draw Hydra.|
|Jun-28-05|| ||catfriend: There is no such mystical thing as understanding the position, just some intuition to shorten path and to evaluate a bit more precisely...|
And more time is <exactly> what Hydra needs to become ubeatable! One ply isn't bad, but I think you're wrong, though. The algorythms being used are genetical, bad lines just die away, so after enough time Hydra concentrates on the best lines. Given <amost unlimited time>, it will approach perfection.
|Jul-01-05|| ||lostemperor: <catfriend> call it what you want. I call it understanding since that is the most appropriate word I can think of.|
Again giving Hydra more time won't matter much but it would make a huge difference for man. The reason why is that it cannot calculate more than say 20 plys (moves for black or white or 1/2 move) deep or it would take years to do so.
From my post at the Kibitzer's Cafe Aug-31-04 <The numbers of different games of 40 moves has been estimated at 25 x 10 to the power of 115 (thats a 1 with 116 0 behind). which is much greater than the number of electrons in the universe estimated at 10 to the power of 79>. Here from a post of <cheski: "The possible moves in a chessgame exceed the number of atoms in the known Universe. (10 to the power of 120th) Even if a computer could calculate one 40-move game per microsecond (and Hydra can't), it would take the program10 to the 90th years to make its first move. Not your ideal spectator sport.">
Take those two correspondence games for instance. Nickel used computer's assistance but a (top) human mind don't need it. Hydra lost them both. I've read that a move should be made within 10 days. If Hydra thought for 10 days about a move, I doubt it would have done much better. Man on the other hand will be incredibly much better if it can analyse the position for 10 days ...
|Jul-01-05|| ||Chesschatology: <lostemperor> Nice point- I fully agree with you. Human's are still the best analysts, just not the best OTB players.|
|Jul-02-05|| ||catfriend: Why did Nickel use them, though, if he didn't need them? Obviously, most his lines were checked and re-checked by the silicon beasts.|
All my point in several posts was that more time <would> help Hydra, and repeating the info about the possible chess moves is useless, as (see last post)<The algorythms being used are genetical, bad lines just die away, so after enough time Hydra concentrates on the best lines.> So that time would be used very effectively. Just in case the second repetition wasn't more effective than the first one, another rephrasing of the same idea: Hydra doesn't calculate all possible moves. Very soon, all it thinks about is the best lines. Given more time, more precise lines are found, and it concentrates only on them. So unlimited tome <does> help it a <lot>.
If we're into quoting, Kasparov said about 28.Ra8!! which didn't occure in his game with Topalov, 99, that it was found probably by help of computer, as Topalov alone couldn't find it without software. So, I"ll trust Kaspy...
<Chesschatalogy> See the quote by Kasparov I've mentioned. The best analysts today sit in front of a computer...
|Jul-02-05|| ||lostemperor: <catfriend> I also trust Kasparov (most of the time).|
“I don’t feel that computers are better than the top humans today…machines that are demonstrably better than Deep Blue are not yet superior to human players”. Garry Kasparov in a post-retirement interview.
|Jul-02-05|| ||offramp: I actually agree with that comment by Kasparov.
What that means, and this is very important, is that a human brain is at least equivalent to a computer such as Hydra.
|Jul-02-05|| ||catfriend: <lostemperor> This quotation doesn't relate to the subject we were talking about. It's another subject, whether machines are better now than players (not precisely what we were discussing) and I think that Hydra (which wasn't familiar in it's most powerful form to Kasparov) answered that question. As to analysis and infinite time you have still to counter my arguements.|
|Jul-11-05|| ||patzer2: White seems to hold with 23. Rd2 =.
Playing it out with Fritz 8 @ 14 to 15 depth gives 23. Rd2 Rb4 24. Rxb4 axb4 25. b3 Ra5 26. Kg2 Ra1 27. Bb2 Bxb2 28. Rxb2 Kg7 29. Kf3 Kf6 30. Rd2 Ra3 31. Rd3 h6 32. h4 Ra2 33. Ke3 Kf5 34. Rd4 g5 35. g4+ Ke5 36. Re4+ Kf6 37. hxg5+ hxg5 38. f4 Rb2 39. fxg5+
Kxg5 40. Rxe7 Rxb3+ 41. Kd4 Rg3 42. Rxf7 Rxg4+ 43. Kd3 Rg3+ 44. Kc2 Rc3+ 45. Kb2 Rxc4 46. Rd7 Kf6 47. Kb3 Rc8 48. Rxd6+ Ke5 49. Rb6 Kxd5 50. Rb5+ Kc6 51. Rxb4 =.
|Aug-06-05|| ||Joshka: So 33.Re6 can hold?|
|Dec-26-05|| ||mr travelchess: I think that Adams could draw here after 51. Kh3!! Ke3
It must be a draw.
|Dec-26-05|| ||THE pawn: <mr travelchess> first of all, Kh3 is the only logical move, so it doesn't even deserve a single ! mark.|
it's not a draw because after 51.Kh3 follows Rxh5+, not Ke3:
Kh3 Rxh5+ Kg4 Rf5 Rxg6 Rf7 Re6 and black is on for the win
Let's say he plays Ke3:
Kh3 Ke3 Rxg6 Rxh5+ Kg4 Re5 Rc6 d4 Rc1 Rd5 Re1+ ( Ra1 is to be considered here) Kf2 Rd1 d3 Kxf4 d2 Ke4 Ke2 Rh1 d1=Q and the game is over for white.
So it's not a draw. I don't believe Adams would have let 10,000$ go away by miscalculating a possible draw!
|Dec-26-05|| ||THE pawn: oh and If 51. Kh3 Ke3 Rc3+ then follows this:
51.Kh3 Ke3 Rc3+ Ke2 (not Kd2 because of Rc6) Kg4 d4 Rc7 gxh5+ Kxf4 h4 Rc2+ Kd3 Rh2 Kc3 Rh1
and it's even worse then the first two lines I gave above.
There might be some flaws because I didn't use a comp. for this, maybe someone could back me up on this with fritz or shredder.
|Apr-27-06|| ||rook h3: as a human i will always hope that a human chessplayer can put up a fight against a computer but i think in a few years, computers will destroy humans with black and white, and we won't even be able to draw with white. the biggest plus though, is that we get a better glimpse at "perfect" chess playing. im sure we'll learn a thing or two about chess that we didn't know before through computer chess.|
|Jul-08-06|| ||spirit: <lostemperor> i agree with you 100% me said it before and me still stand by it: the strongest human will beat the strongest computer if given infinite time.But this is not far fetched anyway...we taught them in the first place!|
|Apr-14-07|| ||silencerius: I have some trouble seeing what happens if white played 51. Kg4 or 51. Kg2 or 51. Kf2 for that matter. Anyone care to help out? In other words did GM Adams quit because he saw its hopeless or was there something in the move 50..f4 that made him quit?|
|Apr-15-07|| ||kapabl: <silencerius> <was there something in the move 50..f4 that made him quit?> If the King comes up Black plays g5+ or gxh5+ winning the rook. If Kh3 then Rxh5+. So the King has to go back. After that, it looks hopeless. Black can play Ke3 snd advance the passed pawn or Rxh5 and Rh2+ forcing the King further back depending on what White plays.|
|Jul-28-07|| ||unixfanatic: It's a good thing Kasparov wasn't playing Hydra - I think he'd have been even more angry afterwards!|
|May-23-08|| ||gambitking: Why 25. K-KB1?
Why not 25. BxB ch. K x B 26. R-Q2 or 26. R-QR2 to protect the pawn?
The Gambit King
< Earlier Kibitzing · PAGE 54 OF 54 ·