Author: Robert Hyatt
Date: 21:54:54 07/27/99
Go up one level in this thread
On July 27, 1999 at 19:15:04, José de Jesús García Ruvalcaba wrote: >On July 27, 1999 at 18:50:08, KarinsDad wrote: > >>On July 27, 1999 at 18:25:20, Dave Gomboc wrote: >> >>>On July 27, 1999 at 17:28:48, KarinsDad wrote: >>> >>>>On July 27, 1999 at 14:07:22, Dave Gomboc wrote: >>>> >>>>>Last night I reached the following position as White in the blitz portion of a >>>>>game/30: >>>>> >>>>>White: Kg2, Qd5, pawns on b3, f2, g3, h4. >>>>>Black: Kf6, Qe5, pawns on b4, g7, h7. >>>>>White to move. >>>>> >>>>>I played 43. Qc6+, and my opponent blundered with 43...Qe6?? I went on to win >>>>>the pawn ending after 44. Qxe6+ Kxe6 45. Kf3. >>>> >>>>Why was this a blunder? There are no good moves. What better move did black >>>>have? >>>> >>>>43 ... Ke7 44. Qb7+ any 45. Qxb4 winning a pawn >>>>43 ... Kf7 44. Qb7+ Qe7 (to protect the b pawn) 45. Qxe7 and black's king is >>>>even further back than with what was played >>>>43 ... Kf4 44. Qf3+ Kg3 45. Qg4+ any 46. Qxb4 >>>>43 ... Kf4 44. Qf3+ Kg4 45. h4+ Kg3 46. Qg4+ any 47. Qxb4 >>>>43 ... Kf4 44. Qf3+ Kg4 45. h4+ Kh3 46. g4 and black cannot take the b pawn or >>>>47. Qf4+ Kg3 48. h5++ or Kh4 48. Qg5++ >>>>43 ... Kg4 44. h4+ and a variety of other moves which lead to the same capturing >>>>of the b pawn or mate with the same themes as above >>>> >>>>And eventually, no matter how black plays, once white wins the b pawn, all white >>>>has to do is place the queen on b7 and push the b pawn. If black tries to check >>>>white first, white plays Kh3, interposes his queen on any other check, puts his >>>>king back on g2, eventually plays Qf3, and can push b3 to protect the b pawn. It >>>>will be REAL hard for black to prevent white from trading queens and having a >>>>passed pawn. >>>> >>>>This seems lost no matter what, so how exactly did black blunder here? Not >>>>trading queens does not delay the inevitable much, if at all. >>>> >>>>KarinsDad :) >>> >>>Some of your variations don't make sense. You must have reset your board to the >>>wrong position. Specifically, White has a pawn on h4, and some lines you have >>>Black playing ...Kg4 and White playing h4+. >> >>My apologies. I had the pawn on h2. It doesn't change much except to prevent the >>Kg5 line. I also screwed up and keep placing the king on f4, g4 and g3 instead >>of f5, g5, and g6 (I have a type of dyslexia when I type variations from my >>head). >> >>43 ... Ke7 44. Qb7+ any 45. Qxb4 winning a pawn >>43 ... Kf7 44. Qb7+ Qe7 (to protect the b pawn) 45. Qxe7 and black's king is >>even further back than with what was played >>43 ... Kf5 44. Qf3+ Kg6 45. Qg4+ any 46. Qxb4 >>43 ... Kf5 44. Qf3+ Ke6 45. Qb3+ any 46. Qxb4 >> >>> >>>Your general argument is that White can pick up the b-pawn, and this is true. >>>(That had been my plan until I was presented with an easier win.) After White >>>grabs the b-pawn, he needs to cover b3 and f2 without allowing perpetual check, >>>and slowly work his b-pawn forward. You're right in saying that this also wins. >>> It's not really so much that it takes longer in the terms of "number of moves", >>>it's just that it takes thought. In the game as played, White was able to play >>>on auto-pilot. The win is clear, and Black has zero counterplay. This was >>>important: we only had about a minute or so each at that point. >>> >>>"Blunder" is how I felt about allowing the queen exchange as the game was being >>>played. I'll agree that that is too strong a word when there's plenty of time >>>to look at the position, though. Part of my impression was because Black _had_ >>>his king in the corner earlier, but instead of playing h6, or g6 and h5, he went >>>and played it into the middle. In the game he had time to go back and hide, but >>>by the time we reached the position that I posted, it was already too late for >>>that. >> >>The point was that Black is lost. Crafty realized it was lost and saw Qe6 as the >>best of a series of lousy moves, most of which lose an additional pawn. Saving >>that pawn and preventing perpetual checks is very easy for a computer, so I did >>not understand the term blunder when it seems the best move, unless you want to >>try to swindle a human into a draw or you are playing against a human with low >>time. >> >>I don't see Crafty not picking a swindle move in a lost game as being inferior. >>That's all. >> > > Maybe Crafty needs to study «The art of the middlegame» by Keres and Kotov. It >has an excellent section onf «How to defend difficult positions» by Keres. >Highly recommended. He does not care if a position is theoretically lost or not, >the goal is to make the opponent's task of winning as hard as posible. > Includes in-depth analysis of a critical position of a game between Capablanca >and Rubinstein (St. Petersburg 1914, I think), a Queen ending in which >Rubinstein was a pawn up. Capablanca defends tenaciously and draws the game. this is a _lot_ easier said than done. IE which move is the best one to try? The one that leads to the deepest lost (mate)? The one that leads to a shallower loss if your opponent plays perfectly, but which might draw if he makes a mistake? That is _not_ easy to determine in the alpha/beta framework we all live in. Seems easy. But it is definitely _not_. The >critical position is one in which all the critics said Rubinstein let the win >slip away. Keres explains Capablanca's defensive *plan* after the move suggested >by the critics, and then backs it with tons of variations. > I think the study of these analysis alone make the book worth having. I wonder >how these concepts can be incorporated to chess engines, as yours seems >completely different it might be more natural for you to include them than for >the other chess programmers. >José. > >>> >>>The main point of my post was that Crafty was misevaluating the position after >>>the exchange of queens. The static eval is about 1.3, and it probably needs to >>>be something like 4 (higher than an extra minor piece, anyway.) The tricky part >>>to doing that is to make sure you are not assigning scores like that to pawn >>>endings that aren't clearly winning at the same time. >>> >>>Dave >> >>KarinsDad :)
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.