Difficult puzzle: white square geometry

Sjugirov-Blomqvist

This puzzle, with white to play, stumped me. (source Chess Today, edition 4498). It is one of the annoying type of puzzles where I know in practice I would never even have considered the move.

Pause now, and attempt to solve it, before reading further

Worse than that, I actually thought I had solved it, and actually thought it was fairly staightforward, but when I looked at the answer, and found it was entirely different from mine, alas, my ability to calculate several moves ahead was shown by Houdini3 to be what it is: poor.

I thought 1 Re5+ won, my line being 1…Be5+ 2 Rd8+ Kf7 3 Qd5+ Qe6 (Re6 4 Qe7+ mates), swap the queens off, take the rh8, and…I had 'mis-visualised' the end position, not realising that black still had a bishop on, on e5, so that after 5…Bb2 black is no worse- I say this because I suspected black was better in the end position, but Houdini says it is virtually equal.

To end to my tale of woe, Houdini shows how my line could have been a draw, by perpetual. 3 Rd7+ and the king has nowhere to run: if it goes to e6, then Qd5 mates, and if it runs via g8 to h7, then the pretty Qh1+ mates. So, draw, but again I know that if I had played 1 Re5+ in practice, I doubt I would have noticed I need to bail out. Perhaps I am being too harsh on myself though.

Solution

Sjugirov played 1 Qb3 which is surprisingly effective. I think it has two points. The first, which tallies with my line, is to strengthen the threat of Re5+ since the Q now blocks the king's escape square, f7. The game continuation was 1…Qg5 2 Qb7 Qh5 3 Qc6+ Kf7 4 Rd6 1-0: a quite nice echo is that in the end position, the Qc6 defends the same square, h1, as earlier.

Victory on the white squares.

Posted with Blogsy