Last week’s puzzle was taken from an article in the latest edition of New in Chess in which Noam Manella looks at moves discovered by engines which are hard for humans to see, but, once seen, are easy to understand.

A very interesting topic, I think, and fertile material for puzzles.

This is from an analysis of the game between Mark Paragua and Das Debashis at the 2012 Parsvnath Open in New Delhi.

Black’s a rook, bishop and knight ahead, but has to meet the threat of mate on the move. Some GMs, according to Manella, suggested 24… Re1+, missing 25. Bd1#. Engines will immediately come up with the correct solution: the amazing 24… Qg4!!, placing the queen on a square where it can be taken for free in four ways.

25. Rxg4?? Re1+ 26. Kd2 Bb4# is rather attractive (bonus points for spotting that one!)

25. Qxg4? threatens mate but the simple 25… Rg8 turns the tables.

25. hxg4? threatens Qf6+, again with a mating attack, but is most simply met by 25… Bf4+ 26. Kd1 Bg5.

Which leaves 25. Bxg4+! when Kg6? gets mated by a north-westerly bishop move, but 25… Kg7! makes a bid for freedom. Now 26. Bf5+ Bg6 27. Bxg6 fxg6 28. Qg4 seems to lead to a draw: for instance 28… Kh8 29. Qh4+ Kg7 30. Qg4 with repetition, or, here, 29. Qxg6 Bf4+ and 30… Rf8 holds. So more bonus points for answering the second part of the question correctly: the position offers equal chances with best play.

You might also note that 25. Qf6+? Qg6! will leave Black with too much material for the queen.

All quite extraordinary!

Perhaps an easier question for problemists, who have learnt to look for this sort of idea, than practical players. Only the other day I saw a mate in 2 where the key was a similar fourfold sacrifice.