The Chess Mind

Bright and early Tuesday morning, I presented a position from IM Dave Vigorito's last-round U.S. Open win over IM Andranik Matikozian. Here's the position, with Black to move and win (brilliantly):



I requested that commentators not use computers and to bear in mind that even if they found the solution, to bear in mind that they, unlike Vigorito, were working without pressure and in the knowledge that there was something special to be found.

Much to my surprise, a couple of commentators wrote in as though the position were just trivial. Tom suggests that "a fairly straightforward 'King Hunt'" should do the job, while Alex Herrera gives the impression that as long as Vigorito doesn't stick his finger in a light socket, he could find a winning move by consulting with a chimpanzee (i.e. Black has so many moves win that picking at random should suffice).

I disagree with them both. The king hunt idea is obvious - Tom's right there - but how to finish it is not - neither Vigorito nor a number of other strong, even titled players spotted the coup de grace. In response to Alex, I think only three or four moves are winning, not 20 or so. Black has the advantage here, and moves that don't commit hara-kiri maintain that advantage. But they aren't "winning" in a non-computer sense. And of those four moves, only two are really decisive. The other two are useful and make progress, but still leave work to be done, especially in a pressure situation with the clock ticking away. And finally, one of those moves only works with the deep idea, while the other works best with it!

I'll discuss the four main candidates in reverse order of strength. First, 45...Rb2, engaging in a bit of overkill on the c-pawn, is a reasonable move if there's nothing better, and Black is computer-winning afterwards. (By this I mean that the silicon oracle bequeaths the desired '-+' symbol on the position, even if only by 1/100 of a pawn and no matter how messy it still is from a human perspective.) Black's position is certainly more comfortable, and while I'd expect Vigorito to outplay his opponent here for the full point, White's not dead yet.

Option number two is the more direct, routine 45...Bxc2. It looks strong, and it is - but you can only take credit for finding this if you calculated all of the following: 46.Rxc2 Rxc2+ 47.Nd2 Qb5+ 48.Kf2 Rxd2+ 49.Bxd2 Qf1+ 50.Kg3 e4+ 51.Qxd6 Qf3+ 52.Kh2 Qh5+ 53.Kg3 Rb3+ 54.Kf2 Qf3+ 55.Kg1 Qd1+ 56.Kh2 Qh5+ 57.Kg1 Rb1+. Black can invert his 47th and 48th moves, but otherwise, every single move is forced if Black desires more than equality. It's not trivial, folks.

Option number three was chosen by Vigorito: 45...Rxc2. White is losing badly on "normal" grounds unless he plays the obvious 46.Rxc2 - which he did. Vigorito played the equally obvious 46...Bxc2, but after 47.Nd2 White would have an edge. (The game concluded 47.Bc5? Rxb3? (47...Bb8!) 48.Bxd6= Qa7 49.Bc5?? (49.Qg7+ is best, but the resulting dead draw meant $20 instead of a possible $860. Now White is better on all continuations...but one:) 49...Bd3+! 50.Kd2 Qa2+ 0-1

After 45...Rxc2 46.Rxc2, Bxc2 was a serious error, and Vigorito saw the right move - which is also the right move (and idea) in the original position - ...Qb5+. Let's take a look at this from the original position and see how it goes.

45...Qb5+ 46.Kf2 (46.c4 and 46.Rd3 also get mated eventually - I leave that as an exercise to the readers) 46...Qf1+ 47.Kg3 e4+ 48.Kh4! (46.Qxd6 is easy to beat: 46...Qxf3+ 47.Kh4 Qh5+ 48.Kg3 Qg4+ 49.Kf2 Qf3#)



Okay, now what? Sure, Black has lots of checks, but no obvious mate and no time for quiet moves, as White threatens Qg7#. One nice try is 48...Bg3+, when 49.Kxg3 walks into 49...Qxf3+ 50.Kh4 Qh3+ 51.Kg5 Qxh6# and 49.Rxg3+ Qh1+ and ...Qxh6#. Vigorito saw all this, but also saw White's best: 49.Kg5! - White wins.

Next, let's try 48...Qh1+ (also considered by Vigorito). Now White plays 49.Rh2! Bg3+ 50.Kg5! Rc7! 51.Rxh1 Rxh1 (threatening 52...Rh5#) 52.Qf8 Be5 (renewing the mate threat) 53.Qxf5 gxf5 54.Kxf5 with a murky position that seems to be in White's favor.

Third, there's 48...g5+, which was my own second thought after 48...Bg3+. (Note that this solution doesn't work after the preliminary 45...Rxc2 46.Rxc2.) After the sequence 49.Kxg5 Rg8+ 50.Kxf5 Qxf3+ 51.Ke6 Re8+ 52.Kd7 Qxf6 53.Rg7+ Kh8 54.Kxe8 we have a position that's a bit confusing in its own right (White is threatening moves like Rf2 and Bd4), and would need to be assessed properly back on move 45. As it turns out, Black is winning with 54...Qc3, but this is no trivial task without Fritz or his brothers.

Finally, there's the clear-cut and especially beautiful solution starting with 48...Qe1+!. 49.Bf2 offers no resistance due to 49...Qh1+ and ...Qh5 or ...Qh6#, but it's not obvious what has Black gained after 50.Rgf2 (or 50.Rdf2) (indeed, Vigorito got this far in his analysis, but failed to clear the last hurdle).



Answer: 50...Qxf2+!! It's a very easy move to miss, as the follow-up to 51.Rxf2 is 51...Rh1+, when Black has sacked the queen to give a check that could have been given, for free, by a stronger piece. It's obvious why it works (52.Kg5 Rh5#) and why, in retrospect, it needed to be done, but it's very easy to miss - especially when it needed to be found on move 45, when there are so many other interesting possibilities, and no one's telling you that you have a forced, brilliant win!

Note: Black can also win with 50...Rc7! (though like 49...g5+, this is unavailable to Black after 45...Rxc2 46.Rxc2 Qb5+ [47.Kf2 Qf1+ 48.Kg3 e4+ 49.Kh4 Qe1+ 50.Rgf2]). This isn't as quick or attractive as 50...Qxf2+!!, but even 50...Rc7! 51.Qxd6 Qxe3! 52.Qxc7+ Kxh6-+ is a challenging find back on move 45.

For those of you who found the main line with 50...Qxf2+ - without the help of a chess engine - definitely give yourselves a big pat on the back! If not, well, there's always next time...