|
< Earlier Kibitzing · PAGE 1 OF 2 ·
Later Kibitzing> |
| Nov-26-05 | | Steppenwolf: 39...Bc2?? Did somebody throw Fritz off the table? Black was winning until that stupid move! |
 |
| Nov-26-05 | | Professeur Y: Bc2 is a real blunder which is quite unusual for a computer, but it happens; I guess computers are only human after all, as the cliché goes. Here is an much worse example by Shredder. Watch how it just forgets to recapture the bishop on b7 when it plays 19...Rfd8??: P Lafuente vs Shredder, 2005 |
|
| Nov-26-05 | | Professeur Y: Here is the explanation given on Chessbase by David Levy, arbiter of the Bilbao tournament, computer chess expert, and frequent collaborator of the Chessbase site:
<The spectators were expecting Ponomariov to capitulate at any moment, but then fate took a hand in the proceedings. Fritz found a move, 39…Bc2, which was clearly winning but for one thing, it actually lost! This is one of the fascinations of computer chess. When you play through this game, take a look at the forced sequence commencing with Fritz’ …Bc2, and count how many ply there are from there to the position when Matthias Feist resigned on Fritz’ behalf. Then add a few more ply because, in the final position, although Black will inevitably be saddled with a decisive material deficit, at that moment the program still had two minor pieces for a rook. So the depth to which Fritz would have needed to search in the critical variation, in order to realise that …Bc2 was a losing move rather than a winning one, was quite beyond the program’s capability.> I have trouble bying this "explanation", which doesn't really explain anything. Perhaps one has to remember that Chessbase is the publisher of Fritz, and suspect that David Levy is a "friend" of Chessbase, trying to "cover up", or at least to minimize, some sort of defect in the brand new Fritz 9 program. Whichever way you look at it, Bc2 is quite an awful move, and it doesn't require such a great "depth" of analysis to see it; I'm quite skeptical when I hear a computer can't figure this out. Is 17 ply, or let's say 20, to add a few like Levy suggests, too deep for Fritz, especially considering the sequence is forced? Besides, the win seems obvious even to me before the final position. |
|
| Nov-26-05 | | MarvinTsai: <Professeur Y> You are right, computers nowadays don't make such "shallow" mistakes, unless time limit pressure. Maybe we should check how much time left on the clock when 39...Bc2 was played, anyway it'd been near 40 moves! |
|
| Nov-26-05 | | DaVinciCodePHI1.618: Black's 24th and 25th moves gained a tempo, and from then on, it didn't look like FRITZ could lose, but it showed that even machines are only human ;) |
|
| Nov-27-05 | | Steppenwolf: Thanks Professor Y, I didn't know that story. Bc2?? just jumps at you when you replay the game. |
|
| Nov-27-05 | | aw1988: To be fair, Junior also picks the same move, so they're not patronizing Fritz. |
|
| Nov-28-05 | | ajile: This is the type of opening you should play against the computers. This is kind of a Colle, Stonewall, London System setup where White provokes the computer to close the position. It's a solid equal position for White. It's well known that computers don't do well in closed strategic positions. The strategy for the human is to basically run the computer out of useful moves. Then the human can build up his pieces behind a pawn chain for example before a sudden breakthrough. Computers thrive in open tactical situations. So playing a Sicilian as Black is basically suicide IMO. When I saw the tournament I immediately noticed this game. It's noteworthy that White had the advantage in the opening and most of the middlegame. |
|
| Nov-28-05 | | wintep: Is Black really better before ..Bc2? What would be a good move instead of Bc2? |
|
| Nov-29-05 | | psmith: <wintep> simply 39...Qxg3+ 40. Kxg3 f5 is good for Black. |
|
| Nov-29-05 | | Radu: <Is 17 ply, or let's say 20, to add a few like Levy suggests, too deep for Fritz, especially considering the sequence is forced?> Fritz was running on a laptop. At about 1000kN/sec you can't expect it to see more than 16 - 18ply ahead and this is with substantial thinking time. 20ply is way too much for such hardware. |
|
| Nov-29-05 | | wintep: <psmith> Yes. Queen exchange, and then f5 followed by 41. h4 with the aim of getting the white King to f4. On exchanging on c4 white has the (big) advantage of its King being close to the action. |
|
| Nov-29-05 | | psmith: <wintep> Is your thought that after 39...Qxg3+ 40. Kxg3 f5 41. h4 is good for White? Fritz (5.32) gives 41... f4+ with big advantage to Black (if 42. exf4 Bc4+ and Bxb3). |
|
| Nov-30-05 | | wintep: <psmith> I see. That is clear. My thought was that the pretty strategic position, (1) placement of the Kings, (2) potential bad bishop against strong knight, ... , might be calculated as drawish, or at least something the computer did not like, which 'forced' to make a tactical error, as there are not so many moves left in that case. (I would like to believe some animal instincts can be assigned to the machine) |
|
Feb-13-06
 | | al wazir: What is the point of 34. Nh5 ? Why not 34. Nxd5 instead? White's pawns at c3 and h3 are dead, but white can get them back after 34...Nbxd5 (34...Qxh3 35. Nxf6+) 35. Bxd5 Nxd5 (35...Qxh3 36. Bc6) 36. Qxd5 Rxc3 (36...Qxh3 37. Qxg5+ Kf8 38. Nb3 Rxc3 39. Rxc3 Rxc3 40. Rc1) 37. Rxc3 Rxc3 38. Qxg5+ Kf8 39. Qd8+, etc. (If black plays Kh8 after white plays Qxg5+, white can force the  trade with Qh5.) |
|
| Feb-13-06 | | ReikiMaster: <al wazir> Why not 34.Nxd5 instead? I think 34...Nbxd5 35.Bxd5 Qxh3 36.Bc6 Qxg4+ or 36...Nxg4 both look good for black. |
|
| Feb-13-06 | | babakova: "Puttin' on the Fritz" real clever... Now I cant get that old Taco song out of my head... |
|
| Feb-13-06 | | EmperorAtahualpa: Nice to see a computer failing! To Ponomariov's credit, the mistake Bc2 was well exploited. To which song is the pun exactly referring to? |
|
| Feb-13-06 | | babakova: <EmperorAtahualpa>
Taco--Puttin' on the Ritz
An old 80´s song. |
|
| Feb-13-06 | | EmperorAtahualpa: Thanks, <babakova>. Guess I was too young to know this. :) Nice pun this is then! |
|
| Feb-13-06 | | karlth: Actually it was Hollywood's biggest star at the time, Clark Gable, who sang "Puttin' on the Ritz" first in 1939. |
|
| Feb-13-06 | | babakova: <karlth> I stand corrected. |
|
| Feb-13-06 | | TerminalK: I'm not a 100% behind stone walling . I wonder what a game between Tal and our generation of computers would have been like? Do you think they would ahve been able to predict his unpredicatableness? |
|
| Feb-13-06 | | Knight13: Too much for Frtiz, isn't it? |
|
| Feb-13-06 | | Bobak Zahmat: I wonder how it is possible that such a great software as Fritz makes mistakes like this. |
|
|
< Earlier Kibitzing · PAGE 1 OF 2 ·
Later Kibitzing> |
