Daily chess puzzle: Check Mate # 95

Another puzzle from Dragoslav Andric’s 1981 book “Matni Udar”.

White to play and mate Black

Marmoroš v unknown, simultaneous, Jerusalem 1926

Find the winning line, including some pretty variations; and find the cook.

Solution

The first move, 1 Rg7! is ‘obvious’, being the only forcing move (or, as Purdy says, “examine all biffs“). But the various defences need refined and different handling.

Then, 1…Kg7 2 Qg5+!! (2 Bh6+ Kh6 3 Qf6+ Kh5[] and Black’s threat of back rank mate (starting Qg1+) means it is only a draw) Rg6 (say: others are similar) 3 Qe7+ Kh8 (3…Kg8 similar) 4 Be5+ and mates.

Or 1….Kh8! 2 Rc7!

2…Qc7 3 Qf6+

And now:

A) 3…Kg8 4 Qe6+ Kf8 (4…Kh8 5 Be5+ 1-0) 5 Bh6+ 1-0; or

B) 3…. Qg7 4 Nf7+ Kg8[] 5 Nh6+ Kh8 6 Be5!! is a very pretty finale.

The cook

So far, so good; and the above are the lines given in the book, and the lines I found when solving the puzzle. But the Stockfish App on my iPad immediately flashed up only +1 in the opening position.

If you haven’t found why, have a look again, before reading on. It is rather deep.

Solution

As so often, mistakes in analysis are caused by automatic exchanges. In the diagram below, Black doesn’t need to play 2…Qc7.

Instead, 2….Qd6! and White is better, but the game goes on. In fact, after the natural line 3 Qd6 Rd6[] 4 Bd6[] Nc6 5 Rb7 is White necessarily much better after say 5..Rd8? At my standard of play, not necessarily so.

Objectively, White might well be much better, even winning: the Nc6 has no stability and if displaced, the b pawn will advance speedily, especially if Black swaps the a pawn for the g pawn, say. But, plenty of mistakes could happen in my practice.

Posted with Blogsy