< Earlier Kibitzing · PAGE 3 OF 3 ·
|Aug-27-14|| ||Gato: Computer gives 33 Rxa7 mate in 21
Black's "best move" would have been 32 ... Rh8f8
|Aug-27-14|| ||Poulsen: I did not solve this - I got lost on 33.Qxe6+ - but could not find a desicive solution after 33.-,Qxe6 34.Rc7+,Kb8 - where should the rook go?|
|Aug-27-14|| ||zanzibar: <Once> Maybe - but you're focusing on the fact that the other line ends in mate, and so is "just as good".|
But perhaps Shirov missed your longer mate too and just saw a skewer tactic gaining Q+N for 2R (and more)...
34.Ra8+ Qxa8 35.Qxe6+ Kb7 36.Qe4+ Ka6 37.Ra4+
click for larger view
The GM knows that's decisive too.
In which case the tactic I saw at first is good enough to solve the puzzle too:
34.Qxe6+ Qxe6 35.Rc7+ Kb8 36.Re7+ Kc8
( 36...Rd6 37.Rxe6 the Black rook being pinned our rook isn't hanging)
37.Rxd8+ Rxd8 38.Rxe6
|Aug-27-14|| ||GufeldsRevenge: Sadly, Rc4 loses to Rd1 followed by Rh1 mate.|
|Aug-27-14|| ||zb2cr: Hi <zanzibar>,
In your reply to <Once>, after 37. Ra4+, Kb5; 38. Qc4#. So the line still is mate, not just the win of decisive material.
|Aug-27-14|| ||Mating Net: Fire on the Board! I miss this version of Shirov and hope he can recapture the magic.|
|Aug-27-14|| ||Penguincw: Not quite the checkmate I was looking for.|
|Aug-27-14|| ||Chess Dad: I didn't get the problem, but here's what I did see. I think it's winning, but it's certainly not a checkmate (yet.)|
34. Qxe6+ Qxe6
35. Rc7+ Kb8 (forced)
36. Re7+ Qd6
|Aug-27-14|| ||awfulhangover: <Poulsen> I went for 34.Qxe6 too, and it wins easily. The rook goes to e7...|
|Aug-27-14|| ||Chess Dad: I checked the 34. Qxe6 line with a chess engine, and it was rated at about +7.8. |
The computer chose a different move for Black after 36. Re7+ than I expected, Kc8, but that's easily taken care of.
Of course, not as good as a mate, but still winning.
|Aug-27-14|| ||Marmot PFL: Figured it had to be 34 Qxe6+ Qxe6 35 Rc7+ Kb8 36 Re7+ Kc8 (Ka8 37 Ra4#) 37 Rxd8+ Rxd8 38 Rxe6. Never even looked for a quicker win.|
Another disaster for black in the b6 Winawer. Always played 6...c5 whenever I played it.
|Aug-27-14|| ||yadasampati: After 34. Ra8+ Qxa8 i prefer 35. Rc4+ (instead of Qxe6+), because i creates a mate in 2|
|Aug-27-14|| ||kevin86: Black will soon find that his king will be skewered to his queen.|
|Aug-27-14|| ||YetAnotherAmateur: The approach I ended up at that I haven't seen discussed yet:|
34. ... Nxd4
35. Rc7+ Kb8 (Qxc7? Qxc7#)
36. Rxc6+ Ka8
37. Qa2+ Kb7
38. Rc7+ Kb8
34. ... Rxd6
35. Rxd6 Rd8 (Qxd6? Ra8#, Qe4/Qd3? Qc7#)
36. Rxc6+ Kb8
34. ... Qxd6
34. ... Rd7
35. Qxe6 Rhd8
36. Rc4 Qxc4
37. Qxc4+ Rc7
34. ... Rd7
35. Qxe6 Rh6
36. Qe8+ Rd8
Any other lines I'm missing here?
The key principle I'm working from here: Right now, black's queen is required to cover a8, b7, c7, and e6, which means she can't move without a devastating attack being opened up somewhere.
|Aug-27-14|| ||zanzibar: <zb2cr> Yes, I agree, that's the reason I use the "(and more)".|
But we're speculating where in the chain Shirov stopped calculating - maybe saying, "OK this wins" rather than "OK this mates".
My contention, or rather speculation, is that he just saw a winning line, and maybe not a mate.
All we know for sure is he missed the shortest mate.
<Once> is basically saying a mate is a mate. I'm almost saying the same, a win is a win. A GM doesn't need to always find the "best" move, as demonstrated in this game.
(And no, I'm not gleefully saying I saw more in the position than Shirov, because I didn't. But I do think it interesting that he missed the simpler mate.)
It would be nice to know a GM's calculation tree - but they rarely publish the lines if they missed the best line.
As far as tactical problems go, this one has several good ways to win. I always treat them as solved when one finds the "best" solution - here, the shortest mate.
<ChessTempo> would force you to find it, and would give Shirov's move as an alternative.
I also think it typically would fail my move, 34.Qxe6_, as a non-mate. Actually, it would probably not publish the position, since there is a winning, but non-mating, line.
|Aug-27-14|| ||patzer2: Shirov seems to be in control the entire game in crushing this French defense. |
After 7. Qg4! Ng6 (7...0-0 ) 8. h4! he gains a clear advantage in space and development and slowly but surely increases the advantage until it's decisive.
Very instructive game, with some good ideas for dealing with the French defense from White's perspective.
|Aug-27-14|| ||gawain: Like many other kibitzers I immediately saw the mate via 35 Rc4+. Must be because I was focused on deflecting the knight away from the defense of the c7 square. Never considered simply taking the annoying knight!|
|Aug-27-14|| ||patzer2: Shirov's 17. Nf4! to is a particularly clever positional move, ignoring the attack on the Bishop to improve the position by forcing the exchange of Black's strongest piece in White's favor.|
For those who saw <Ruzon>'s ABCDE post at J D Turner vs I Ponter, 2004 about the five ways (move away, block, capture, defend, something else) to defend against an attack on a piece, this move (17. Nf4!) is "something Else."
|Aug-27-14|| ||jdc2: <gofer: jdc2 Isn't your position mate in 5?>|
click for larger view
Analysis by Stockfish 190614 64 SSE4.2:
1. (#6): 1.Nxd5 Rd7 2.Rxd7 Qb7 3.Nxb6+ Qxb6 4.Rc1+ Nc4 5.Rxc4+ Qc5 6.Rxc5#
|Aug-27-14|| ||Ferferi: It's been a known fact that spotting tactics in real games is far more difficult that when they are presented as puzzles. If it's a puzzle, you already know there's something to look for.|
There's a flip side to that, though. I "saw" 34. Rc4, where taking the rook leads to mate, and this being just a 2-star puzzle, I decided that that must be the answer without much of a second thought. It must've been exactly what black was counting on, as it runs into one of those horrible back-rank punches that I'm so much used to receiving...
|Aug-27-14|| ||ChessGeezer: Does 35 Rc4+ win too? It looks easier than Qxe6 to me. I must be missing something.|
|Aug-27-14|| ||morfishine: White has the winning deflection <34.Ra8+>|
I followed 34...Qxa8 <35.Qxe6+> Kb7 36.Qe7+
I sure wish I had seen the simpler 35.Rc4+
|Aug-27-14|| ||nateinstein: <ChessGeezer: Does 35 Rc4+ win too? It looks easier than Qxe6 to me. I must be missing something.> The same thing I almost missed...Rd1+ leading to mate for black :). The queen is pinned, but not the rook. The rook goes to h1 once king moves up.|
|Aug-27-14|| ||TheBish: Shirov vs S Kristjansson, 2004|
White to play (34.?) "Medium/Easy"
I noticed that if the black queen could be diverted from her guard of a8, then Ra8# would be the result. But 34. Rc4?? (hoping for 34...Qxc4 35. Ra8#) would lose quickly to 34...Rd1+ 35. Kh2 Rh1#. (Always look for checks!)
As is often the case in tactics, if one plan doesn't work, try changing the move order! In this case, it also consists of making a more forcing first move -- a check instead of an attack on the queen.
34. Ra8+! Qxa8 35. Rc4+ Nc5 36. Qc7#.
|Aug-27-14|| ||yadasampati: <morfishine> Indeed, that is what i also said before: 35.Rc4+ is better than 35. Qxe6+, because it creates a mate in 2|
< Earlier Kibitzing · PAGE 3 OF 3 ·