The Chess Mind

A couple of days ago, I presented this difficult ending as a challenge to the reader: White to move and win.



I first came across the position in the 1997/6 issue of New in Chess Magazine, pages 5-6, where the author, one Srdjan Sale gives us an idea of just how difficult it really is. I quote:

I recognized this position is of a special category [of difficulty - DM] and to prove my hypothesis I started a pseudo-scientific experiment by showing it to GMs and IMs. Here is a list of 'celebrities' who, after one hour, lost the bet or simply resigned from the torture: Cvitan, Kozul, Hulak, Cebalo, Vl. Kovacevic, Wahls, Tukmakov, Sax, Lalic, etc., etc. I even got the opportunity to test one of the Top Ten players, i.e. Alexey Shirov...[and] even Shirov had to admit he couldn't solve it!

Later Dreev and Zviagintsev tried some 'intuition method' from the Dvoretsky School but failed. The next day Zviagintsev told me that I had spoilt his peaceful night with this rook ending but at least he finally had come up with the solution.

The only two players who solved the position are GM Dizdar in about twenty minutes and the young Brazilian IM [now a GM - DM] G. Vescovi in about forty minutes.

According to Sale, the position came from the game Sulava-Godena, Montecatini Terme Open 1994, but the game score in the Mega Database 2005 doesn't ever reach that position - in fact, the f-pawn never made it past f6. My guess is that either the game degenerated in a time scramble and the remaining moves weren't recorded, or else they happened upon this position in post-game analysis. I find the first explanation implausible, as the final position given in the database is a trivial draw that couldn't plausibly have reached the position in the game, so I'm going to assume the latter unless a plausible counter-story comes in.

The position can also be found in a 1953 endgame study by N. Kopaev, which started with the Black rook on c2 and with Black to move: 1...Rc8+.

Enough by way of background: let's solve the thing! If you're ready to see the solution, click below.

Here's a tough little nut for the aspiring reader to crack; the solution will be given in a couple of days.



White to move (obviously) and win.

When I do my ChessBase shows each Monday night, I do my best to offer comments that will be instructive to a wide range of spectators, but it's not possible to answer every question in advance or even during the show (in part because of the time delay, and in part because it will generally mess up the program's narrative flow - especially for those who listen afterwards in the archives).

In this post, then, I will answer two recent questions that weren't answered during the programs themselves. The first pertains to this past Monday's show, on the crazy Gabriel-Korchnoi game presented in a previous post.

Here's the position in question:



White is down a pawn and (much more importantly) severely bottled up by the f5-d3 pawn chain. He needs to decide whether to work around the pawns or chip away at them, and Gabriel chose the first option with 15.Na4. This was criticized by Korchnoi and other commentators, who thought that option two, by means of the dramatic-looking 15.g4, was the better try. In my view, 15.Na4 was probably okay - I think the key error came later - but are these the only two options?

Not according to one spectator, who offered a radical suggestion: why not 15.Nxe4? White gets to collect all three of the horrible pawns in return for the piece, after which his pseudo-bishop on f1 is reordained a chess piece while the Black kingside looks a little airy.

It's a very nice idea in theory, and sometimes that sort of sac works like gangbusters (for those who know the game, see Yudasin-Monokroussos for a dream-come-true version of a piece-for-two-pawns-and-a-big-pawn-center sac). Here, unfortunately, Black is well on top after 15...fxe4 16.Bxc5 Qxc5 17.Qxe4 Re8 18.Qxd3 Nf6 followed by ...Kf7, and Black is winning. His king is safe, his pieces are active, and White's central pawns are not at all likely to have the same effect Black's pawns had before the sac.

A second comment came via email to the old blog. [N.B. Readers, please send your questions to me here - just click on "Contact" on the right sidebar, and you'll be able to write me in that way.] Lee Prince asked about a game I did long, long ago, a Berlin Defense between Harmonist and the "Praeceptor Germaniae", Siegbert Tarrasch.

After the moves

Harmonist,Max - Tarrasch,Siegbert [C67]
DSB-06.Kongress Breslau (3), 1889

1.e4 e5 2.Nf3 Nc6 3.Bb5 Nf6 4.0-0 Nxe4 5.d4 Nd6 6.Bxc6 dxc6 7.dxe5 Nf5 8.Qxd8+ Kxd8 9.Bg5+ Ke8 10.Nc3 h6 11.Bf4 Be6 12.Rad1 Rd8 13.Ne4 c5 14.Rxd8+ Kxd8 15.Rd1+ Kc8 16.h3 b6 17.Kf1 Be7 18.a3 Rd8 19.Rxd8+ Kxd8 20.c3 Bd5 21.Nfd2 Kd7 22.Ke2 g5 23.Bh2 Nh4 24.g3 Ng6 25.f4 Ke6 26.Ke3 c4 27.Nf3 gxf4+ 28.gxf4 c5 29.Ng3 Nh4 30.Nxh4 Bxh4 31.Ne4 Be7 32.Bg1 Bc6 33.Bf2 Bd7 34.Bg3 Kd5 35.Nf2 h5 36.Kf3 Bf5 37.Ke3 b5 38.Kf3 a5 39.Ke3 b4

we reach this position:



The game continued 40.Kf3 Kc6 41.axb4 cxb4 42.cxb4 axb4 43.Ne4 Kd5 44.Nd6 Bxd6 45.cxd6 c3 46.bxc3 b3 and White resigned.

But perhaps White can improve? Prince asks, "What about (40.h4)? I'm a total novice but I let Fritz crunch on it and it seems ok for white. He can then get his knight in the game or force black to trade off the light square bishop. What do you think?"

In some positions, little finesses can make a difference, especially when the question of a win or a draw comes down to a tempo or two. In other positions, the static features predominate, and one side's chronic weaknesses will doom him - a tempo here and there won't make a difference.

This position is of the latter sort. Black is breaking through on the queenside, and White's options are essentially limited to determining whether Black's king comes in via d5 or b5. Here is a typical line:

40.h4 Kc6 41.Ne4 Kb5 42.axb4 (Forcing an immediate exchange with 42.Nd6+ Bxd6 43.exd6 allows Black to grab loose queenside pawns and win after bxc3 44.bxc3 Ka4 45.Kd2 Kxa3 46.Bf2 a4 47.Kc1 Kb3 48.d7 Bxd7 49.Kb1 Kxc3 50.Bxc5 Kb3 51.Bb6 Bf5+ 52.Ka1 c3 53.Be3 Kc2-+) 42...cxb4 43.cxb4 axb4 44.Nd6+ Bxd6 45.exd6 and now the same tactic from the game does the trick: c3! 46.b3 (Trying to save material in the short term with 46.bxc3 lets Black queen the b-pawn with b3 47.d7 Bxd7 48.Kd2 Ka4 49.Kc1 Ka3-+) 46...Kc6 47.Ke2 Be6 48.Kd3 Bxb3 49.f5 c2 50.Bf4 f6 51.Bd2 Kxd6 52.Bxb4+ Ke5 53.Kd2 Kxf5 54.Bc3 Ke6 55.Bb2 Ba4 56.Ke3 Kd5 57.Kd2 f5 58.Ke3 Kc4 59.Bc1 Bd7 60.Ke2 Kb3 followed by ...Ka2-b1, winning.

The lines are long and Fritz won't just "get it" right away (though it might find the lines one move at a time), and perhaps there are other defensive attempts White might try. A bit of patience and trial and error, though, and the solutions will come.

Happy analysis!