The Chess Mind

White to move and draw:



It's a neat little puzzler, and when you're ready to see the solution, the answer is but a click away.

White to move and draw (obviously); solution tomorrow.

Yesterday, I offered this analysis position based on the game Chigorin-Tarrasch, St. Petersburg (m/18) 1893; it's White to move and draw:



Tarrasch, in his annotations, believed that the variation that led to this position could be won by White after 52.Kg5 Kxa4 53.Kxh5 Kxb5 54.Kg5 a5 55.h5 etc., as the imminent queen on h8 will cover a1 in the nick of time. Now, Tarrasch's annotations to most of this game were excellent, but there are no less than two blunders in the variation above.

The second one is caught by Kasparov: instead of 52...Kxa4??, Black plays 52...f4! and wins. If White takes the h-pawn, the f-pawn queens, and on 53.Kxf4, it is Black who queens first after 53...Kxa4 54.Kg5 Kxb5 55.Kxh5 a5 56.Kg6 a4 etc.

If that's blunder #2, then obviously 52.Kg5 is blunder #1. But what's the right way? The answer is just a click away...

In the process of analyzing and researching previous annotations to the Chigorin-Tarrasch game for this week's ChessBase show, the following analysis position arose:



White to move and draw.

If you don't get it, don't feel bad: neither did Tarrasch or Kasparov in their annotations. Solution tomorrow.

By way of reminder, the problem to be solved looks like this:



It's White to move, and despite his having an extra queen, the threat of ...g1Q - with check, at that! - makes the position a challenge. Of course White can play 1.Ne4, with the idea of meeting 1...g1Q+ with the winning 2.Nf2+, but it's not at all obvious what he can do after 1...Nd3!, as 2.Qxd3 g1Q+ is a draw.

If you haven't worked on it yet, or if you only came up with 1.Ne4 g1Q+? 2.Nf2+, here's your chance to figure it out. If you're ready to see the solution, however, it's just a click away.

Here's your position; it's White to move and win:



The solution will be given tomorrow.

Martin van Essen writes:

Hi Dennis,

I just read the topic "From the Mailbox: How Badly can a Computer Misevaluate?" from some months ago.

I remember once having setup a position in Chessmaster 9000 (on a humble 500 MHz) involving a black h-pawn (h3 or so) and nine black 'wrong bishops' (eight promoted ones). White's lone king at h1 faced an approximate 19 pawns deficit according to Chessmaster. I'm curious what other programs have to say about this.

My computer is more powerful than yours, but that didn't matter to Fritz, Shredder or Rybka, which gave evaluations ranging from +24 to +34! Indeed, it wouldn't matter if one were running the position on Hydra or Deeper Blue: either the software "gets it" or it doesn't. The hardware problem is this: the game could continue for almost 350 moves just taking into account the process of getting rid of the promoted bishops. So until computers are a lot more powerful than they are now, the only way for chess engines to get it right is by programmers creating a specific rule that 1-8 wrong-colored bishops draw, provided the defending king can reach the queening square.

The Chess Cafe has a new endgame study up each week, but the current entry caught my eye as a particularly elegant, bite-sized problem of recent vintage:



L. Falk, Schacknytt 1990; White to move and win.

You can find the solution at the bottom of the page here - but don't check until you're sure you've solved it!

Please read the previous post first. When you have done so, and have decided that you're ready to see where Hansen went astray (or are checking your own conclusions), you'll be ready to click here for the answer.