A) 44... e5+ 45.Kg5 seems to worsen black's position.

B) 44... g5+

B.1) 45.hxg5 Kg6 46.g4 h4 and black wins.

B.2) 45.Kxg5

B.2.a) 45... Rg8+ 46.Kf4 Kg6 47.g4 doesn't look satisfactory for black.

B.2.b) 45... Re5+ 46.Kf4 Rf5+ 47.Ke3 d4+ and white loses a rook.

In was looking for 44...e5+ and ended somewhere in nowhere. Afterwards I had a closer look at 44...Kh6. I was also looking shortly for 44...g5, but I found no convincing solution.

1) 44...e5+ 45. Kg5 and there seems no way to proceed.

2) 44...g5+! and now

2a) 45. hxg5 (occupying the 1 safe square the king could run to) e5# and

2b)45. Kxg5 Rg8+ 46. Kf4 and now the king is stalemated so 46...e5#!

I was fixated on ...e5+, and I noticed that 44...Kh6 threatened mate. White has no choice but 45.g4 to create an escape at g3. Now I have 45...e5+ 46.Kg3. So forcing! I made his king retreat! My dangerous central pawns were advancing! I was certain that I had the right idea!

If white plays g4 on move 44 it looks OK for him.

44 g4 guards the crucial f5 square and creates the escape square on g3.

click for larger view

44...g5+

White is now faced with a choice: 45. hxg5 or 45. Kxg5

A. 45. hxg5 Kg6
Black now threatens mate next turn with ...e5 and so white has no time for a move like 46. R5b7. A desperate move such as 46. Re7 does not work either. Black can simply capture the e7 rook and again mate is threatened with ...e5 (e.g., 46. Re7 Rxe7; 47. fxe7 e5#). The only move for white looks to be:

46. g4 h4
Mate is now threatened once again with ...e5 and this time white cannot stop it without giving up a rook (e.g., 47. Rxd5). But it is quite hopeless since black can win that rook (47 ...exd5) with an easily won position. Afterall, black has a king that nicely protects the f7 pawn and has a passed pawn on the h-file (another benefit of having moved 46 ...h4) that looks ready to march down the board.

B. 44 ...g5+
45. Kxg5 Re5+
46. Kf4 Rf5+
47. Ke3 d4+
And black wins the b5 rook next move with 48...Rxb5. Once again the black king is on a nice defensive square (g6), the f7 pawn is protected from white's remaining rook, and black should have no trouble with the rest of the game.

Black to play and win.

Material: Even. The White Kf4 is confined to dark squares by the opposing Ps and has 1 legal move, suggesting a mating net revolving around the g5 square. The Re8 requires activation. White threatens Rxf7+, which requires a response.

Candidates (44…): Kh6, e5+, g5+

44…g5+

Most of us try to do it in our heads.

I have found the best way to improve my analytical thinking is (1) keep coming here, (2) discover in very concrete ways how much better most of these people are at analytical thinking than I am, and (3) realize that there isn't too much stopping me from thinking more like they do.

Yes, I saw all of this by just looking at the puzzle diagram (and jotting down the variations in Wordpad and pasting it here) - the most other kibitizers do the same kind of thing.

The answer re. "improvement in analytical thinking" is the same answer as you get to the question: "How do I get to Carnegie Hall?" The answer is: practice. Practicing getting these kinds of puzzles, practice other kinds of positions, play over games, play your own games and then study them, etc...

In Andrew Soltis's book "The Inner Game of Chess," the author cites studies where top players and amateurs were shown various chess positions. The really good players could reconstruct from memory the positions shown to them very quickly and accurately. But these were typical kinds of chess positions. When presented with very atypical positions (i.e., that would never really occur in a real game), the good players were no better than the other players. That is because the positions made no sense - the position of the pieces and pawns had no chess-logical relation to one another - there were no typical patterns and much of chess skill lies in pattern recognition. But pattern recognition is largely an unconscious skill that is acquired over time. We look at the way the pieces are gathered around each other and then bing! something registers consciously in terms of what we should play. Those bings! happen more frequently and more accurately the longer we have been exposed to the game.

Yesterday I thought I had come up with a pretty good initial move but in reality it was a bad blunder. I think these kinds of instances of "chess blindness" happen to everyone - on occasion even strong players. But I think they happen a lot more to a player like me (average tournament strength) in comparison to a better player. Why? Because they are quicker when it comes to calculation, can see further, and are more accurate. And that is really because of practice and hard work on their part over a long period of time (combined to some degree with their natural abilities in the game).

Today's puzzle has a few elements to it that potentially can cause difficulty. The first is that 44 ...g5+ looks counter-intuitive. I mean why give away material? Well for one thing, it gives the black king a nice square to sit on (assuming white plays 45. hxg5) that also tightens the mating net around the white king. But it also diverts the white king away for the e5 square (assuming white takes the pawn with the king) and allows the rook to land on the e5 square with a check. And so while it may be counter-inituitive, especially to players with less experience, ...g4+ is actually quite a forcing move (as most checks are in comparison to other types of moves). White has only two possible replies and after each of those replies, further moves are also fairly forced. So relatively speaking, it does not require all that much in terms of visualization of future moves because we aren't talking about a lot of possibilities here. So in that respect, the puzzle is relatively easy in comparison to other positions in which the number of possible moves and possible responses is great.

The other thing that makes this puzzle tough however, and something I have noticed at other times as well is how the end objective can switch around. Reading the comments, it certainly looks like a lot of others were trying to find way to trap the white king and force mate. Certainly the position of the white king recommends this kind of thinking. But it wasn't until I gave up the idea that it was absolutely the goal to checkmate the white king that I noticed how black could win the b5 rook (in the line where white captures the g5 pawn with the king). In fact I noticed this before even thinking about the line where white captures with the pawn instead. Ironically in that line I initially thought about how the black h-pawn could make it down to h1 and queen before I again had to switch back to the "okay, can we checkmate the king?" question.

My point in mentioning all of this is that there are lots of different ways to approach thinking about a positon. With yesterday's puzzle, lots of people got the answer right away, but I was completely off-track. Today, for whatever reasons, I think I had a bit of an easier time then several others.