< Earlier Kibitzing · PAGE 2 OF 2 ·
|Jan-05-07|| ||The beginner: This was not that hard for friday, eather that or i am starting to get better :-)|
I saw this all way upto Bc5 wining the Queen back, with material and positional advantage.
|Jan-05-07|| ||Fisheremon: 19. Qxd6 looked pretty only in the case Black accepted. Black could improve a bit with 19...Qxg6 still losing.|
19.Ra4 (noted by many participants) seems to be the best move with two main variations:
19...Nd4 20.Nxd4 exd4 21.Rxd4 Qxg6 22.Rg4 Qe8 23.Kd1 Rf7 24.Bc4 d5 25.Bd3 Qe7 26.Rxh7 with fierce attack
19...Rf4 20.Bxf4 exf4 21.Kd1 Ne7 22.gxh7+ Kh8 23.Rxf4 Bxb2 24.Qb4 Bc6 25.Nd4 etc. with winning attack.
There's a variation suggested by <chessmoron> 19.Qd5 Qxd5 (?!) 20.Rxd5 leading to a win too, but if Black plays 19...Nd4, e.g. 20.Qxe6+ Nxe6 21.Bc4 Rf6 22.gxh7+ Kh8 23.Nd2 Nf4 then White faces a serious problem with a winning plan.
|Jan-05-07|| ||Towershield: <think> 20.Nc5 wins easily for white in that line.|
|Jan-05-07|| ||KholdStare: I got up to 23. Bc4. why Bc4 and not Bxg6?|
|Jan-05-07|| ||TrueBlue: Queen sacrifice is totally uncalled for in this postion!!!! White easily wins after gxh7 (no ! needed, quite an obvious move). So I am counting gxh7 as a solution.|
|Jan-05-07|| ||Drudge: what is wrong with line:
? I can't see it, but I am blind, can someone tell me the flaw here?
|Jan-05-07|| ||alphee: I got the first moves rather easily 19.♕xd6 ♕xd6 20.♗c4+ ♖f7 21.♗xf7+ ♔f8 22.♗c5 hxg6 23.♗c4 but missed the 23. ... ♕xc5. Then I looked at ♗xh6 that lead to 19. ♗h6 ♗xh6 (19... ♗h8 20. ♗xf8 ♔xf8 21. ♕h6+ ♔e8 22. ♕xh7 ♕f6 23. g7 ♗xg7 24.
♕g8+ ♔e7 25. ♕d5) 20. ♕xh6 ♖fe8 21. ♕xh7+ ♔f8 22. g7+ ♔f7 23. g8=♕+ ♔f6 24.
♕gg6# *. As the line leading to the check mate is not forced (what Robin Smith calls a "canyon" in this book Modern Chess Analysis <www.gambitbooks.com/books/modanalysis.html>)
I gave up and thinking I didn't get the solution I went to check the game....|
I didn't get yesterday's puzzle at all and missed the 18. ... ♖e3 completly
|Jan-05-07|| ||HelaNubo: <drudge> 20... Kh8 21. Rxh7 mate. Dont worry, it is a kramnik-like mistake;-)|
|Jan-05-07|| ||boogaloo: I'm such a nOOb. I saw 19.Bc4. Aaaaaarghhhh..... Friday puzzles! What should I study to get puzzles like this?|
|Jan-05-07|| ||uuft: 19. Ra5-d5... gave me mate in 7 moves on MacChess (5 sec-level). No need for a queen-sac imo.|
|Jan-05-07|| ||alphee: <TrueBlue> I tried gxh7 too after Qxd6 but was not too convinced.
Reading your comment I asked Fritz to analyse the position without having played the game first and then after the game was played (he "knew" the solution).
Interstingly, at the first pass Qxd6 did not show up but it did the second time.|
1 - Analysis of postion only after it was copied as FEN
click for larger view
Analysis by Fritz 9:
1. (3.41): 1.Ra4 Ne7 2.gxh7+ Kh8 3.Na5 Bc8 4.Bd3 Rb8 5.Nb3 d5
2. (3.20): 1.gxh7+ Kh8 2.Ra4 e4 3.Rxe4 Ne5 4.Rb4 Rab8 5.Nd4 Qe7 6.Ne6 Qxe6
3. (3.09): 1.Qd5 Qxd5 2.Rxd5 Nb4 3.Rxd6 Nxc2+ 4.Kd2
4. (2.50): 1.Rd5 Qxg6 2.Rxd6 Rf6 3.Rxf6 Qxf6
5. ± (1.09): 1.Rh4 Qxg6 2.Rg4 Qf6 3.Bg5 Qf5 4.Bh6
2 - Analysis of the same position after the game was played with Fritz, starting with move 18 for white (I had to go backward).
Analysis by Fritz 9:
1. (4.09): 1.Qxd6 Qxd6 2.Bc4+ Rf7 3.Bxf7+ Kf8 4.Bc5 Rd8 5.Ra4 hxg6 6.Bxg6 Nd4 7.Bxd6+ Rxd6
2. (3.41): 1.Ra4 Ne7 2.gxh7+ Kh8 3.Na5 Bc8 4.Bd3 Rb8 5.Nb3 d5
3. (3.20): 1.gxh7+ Kh8 2.Ra4 e4 3.Rxe4 Ne5 4.Rb4 Rab8 5.Nd4 Qe7 6.Ne6 Qxe6
4. (3.09): 1.Qd5 Qxd5 2.Rxd5 Nb4 3.Rxd6 Nxc2+ 4.Kd2
5. (2.50): 1.Rd5 Qxg6 2.Rxd6 Rf6
This is very much in line with what Robin Smith already stressed in his book MODERN CHESS ANALYSIS.
Chess engine react differently if they are given the position (and only the position) or if they are asked to analyse a position which have been "played" before during the same session and still have in their memory tables!
|Jan-05-07|| ||nimzo knight: I will say I almost got it. I didnt see ..hxg6. I think there are other defences which lead to same loss of exchange.|
|Jan-05-07|| ||YouRang: Got it! The first thing I saw was that black's K+Q were on the same diagonal, so I spent a few moments trying to pin the black Q+K with the Bc4. |
This was to no avail -- except that I noticed that the check from the bishop at c4 by itself is quite strong (it would cost black a rook to avoid mate, i.e. Bc4+ Rf7, then Bxf7+).
So next, I tried to get the black queen out of the way so that I could play Bc4+. Again, there seemed to be no good way to do it except the "ridiculous" 19. Qxd6, losing my queen to 19...Qxd5.
But I looked a little deeper into it, and noticed that after I take the rook (20. Bc4 Rf7 21. Bxf7+), black's K must go to f8, putting it (again) on the same diagonal as the queen -- and this time, the pin (Bc5) works! :-)
|Jan-05-07|| ||kevin86: I went for pinning posibilities on the c4-g8 diagonal-where the pinning blow was on the OTHER diagonal-where the king and queen were both chased. A remarkable finish!!|
|Jan-05-07|| ||MiCrooks: I would have never gotten this in a game, as I would have taken on h7 in heart beat and played the game up two pawns and position. I wouldn't have even considered the Q sac.|
|Jan-05-07|| ||Stonewaller2: <alphee: Chess engines react differently if they are given the position (and only the position) or if they are asked to analyse a position which have been "played" before during the same session and still have in their memory tables!>|
Interesting. I take that to confirm that chessplaying computers are limited by their inputs, that is, the information provided by their human programmers. One might assume (whatever version of) Fritz would find the key eventually, but given the bias of its programmers it doesn't seem to start to search in that direction unless it has been prompted to, in this case by playing the game over first.
In any case it's safe to conclude that Fritz's "reasoning" is not along the lines of <YouRang>'s: seeing a check along one diagonal and concluding it wins a ♖ plus, given an earlier ♕ sac, forces the ♔ onto another diagonal where a pin is possible. Rather, Fritz and its ilk merely search all possible future positions, assign numerical values to each according to some (human-produced) algorithm, then pick the "best" move based on those values.
In this case Fritz's "decision tree" apparently got pruned too soon.
P.S. For the record, I missed it. Saw the pin on the c4-g8 diagonal but no way to make it happen without telegraphing the punch, so went for 18. gh+.
P.P.S. <alphee>, from your "paranoic-critical" pic I assume you'd agree that Chess "beauty must be edible or not at all." ;)
|Jan-05-07|| ||Billy Ray Valentine: Very interesting post, alphee. Thanks for the analysis. I got the puzzle--but had I been in a real game I'm pretty sure I would have played 19. gxh7+ instead.|
|Jan-05-07|| ||TrueBlue: well, I think white's g pawn is very strong. Even without a computer, you can see that white has massive advantage after gxh7. Since black bishop cannot go on h7, it is a real treat and black needs to spend a lot of resources to keep it from becoming a queen. Of course, jmo, no computers used. So I think gxh7 is much stronger than Qxd6, because the later loses the g pawn. Just my oppinion, even if I saw the queen sacrifice in a game, I would still play gh7. I don't believe alphe's computer printout correctly evaluates the strenght of the pawn.|
|Jan-05-07|| ||haha: My answer was 19. ♕c3 Is it good?|
|Jan-05-07|| ||alphee: <TrueBlue> <I don't believe alphe's computer printout correctly evaluates the strenght of the pawn.> that is probably true and the printout was just to show that if Fritz is given JUST THE POSITION, the ♕xd6 does not appear in the evaluation!|
|Jan-05-07|| ||chopin4525: I got it!looking deep shredder analysis i noticed that he fuond as best line 19.Qxd6,Qxg6 20.Qxg6,h7xg6 21.Bc4+,...of course white has always a huge advantage!^__^|
|Jan-05-07|| ||Themofro: I got this one. Having the queen and king lined up on the same diagonal like that is generally speaking risky but there was no way to pin the queen to king. however if the queen could be moved then a check along that same diagonal would be devestating due to the king not being able to retreat to H8 because of the mate that the rook would then deliver. A nice combination and game, good puzzle too.|
|Jan-05-07|| ||YouRang: <haha: My answer was 19. ♕c3 Is it good?>|
I suppose that the point of 19. Qc3 is to support the 20. Bc4 pin, but black of course will see this and probably play 19...Qxg6, removing the pin and the troublesome pawn.
So, is 19. Qc3 good? You tell me: What do you have planned if 19...Qxg6?
|Jan-05-07|| ||ALEXIN: Missed...|
|Jan-06-07|| ||Fisheremon: <alphee: 3. (3.09): 1.Qd5 Qxd5 2.Rxd5 Nb4 3.Rxd6 Nxc2+ 4.Kd2>|
See my message with 1...Nd4 (19th Black's move).
<alphee: 4. (2.50): 1.Rd5 Qxg6 2.Rxd6 Rf6 3.Rxf6 Qxf6>
19...Nb4 20.Rxd6 Nxc2+ 21.Qxc2 Qxd6 22.Bc4+ Bd5 23.Bc5 Qc6 24.Bxf8 Qxc4 25.gxh7+ Kh8 26.Bxg7+ Kxg7 Black could have chances to a draw.
<alphee:1. (3.41): 1.Ra4 Ne7 2.gxh7+ Kh8 3.Na5 Bc8 4.Bd3 Rb8 5.Nb3 d5>
20.Na5 much better almost forcing Black to lose exchange Rf4 still under White's attack (see also my message with 19.Ra4).
My thinking: don't rely entirely on computer analysis (you pointed out it partially with Fritz 9): as an illustration see my analysis of a recent puzzle A Schneider vs Dolmatov, 1983
< Earlier Kibitzing · PAGE 2 OF 2 ·