Test your chess: daily chess puzzle # 102

White to play: is it won or drawn?

AJ Cameron v H Meihuizen 1910

Solution

In the game, white repeated by 1 Qh5+, 2 Qh5+, with a draw: which, if I were white, I would do too, fearing that if I tried to play for a win, then I would risk losing, especially given black has many defences.

Reitstein says that close analysis at the time convinced everyone that white could have won by 1 Bh6! which of course is the natural try. I was put off from 'playing it' (I try to mirror, to the extent I can, the thinking in a game when solving these daily puzzles, though of course it is not fully possible) by 1…Qd7, thinking that it wasn't clear after 2 Bg7+ Kh7[].

However, Stockfish tells me that white is clearly winning (+6) after 3 Qh5+! Kg7 4 Rf1!

At first blush, that this is so seems extraordinary, but by looking at the positron, the reasons became apparent. Black's rook and queen's bishop are out of play; the Bc5 is LPDO; the N can come to e4; if the Bc8 develops on e6, then the e5 pawn drops off…black does indeed seem to be lost.

Back to the main line: 1 Bh6 gh 2 Qe8+ Kg7 3 Rf1:

Reitstein says that at the time, white's threats were thought to be overwhelming, so that the conclusion was that white was winning, until this century (the book was published in 2005) Fritz found 3…Bh3!! with a draw:

4 Qa8 Bg2+ 5 Kg2[] Qg4+ 6 Kh1 Qh3!!

7 Rg1+ and black has a perpetual.

So, Reitstein concludes it was a draw.

But is it? Stockfish (2014) suggests 1 Bh6 gh 2 Rd1!!

The engine doesn't like 2…Qd1+ 3 Nd1 Bg4, defeating it by 4 Qc4, exchanging one of black's bishops, so that it results in Q v R+B where black's king is open, and some of his pawns drop off:1-0. So 2..Qh4 when 3 Qe8+! Kh7 (3…Kg7? 4 Qe5+ and the LPDOS Bc5 drops off due to the fork) 4 Rf1!

There is a ever so subtle but key difference between the above position and the previous line in which Bh3!! saved the day. The difference is not that the black queen is on h4, but that the black king is on h7! (and not on g7). So 4…Bh3 doesn't now work because 5 Qa8 Bg2+ 6 Kg2[] Qg4+ 7 Kh1[] Qh3 which is very similar except that white can now play 8 Rf7+! since the K is not on g7, so the rook can't be captured. 8…Kg6[] 9 Qg8+ Kh5[] 10 Ne4 1-0 (or 9 Qe8).

So, I suspect, after all, it is a win, but by a subtle change of move order. Chess is a truly deep game.

I am not sure what I have written is the last word in 2014, or whether Komodo or another supremely strong engine could find yet more depths. As I finish this posting, Stockfish is pretty confident that 1 Bh6 is winning. As an aside, at first it liked 1 Bg5, but then after 1…Bg4 2 Qc7 Rf8 it assesses the position as 0.0, giving the line 3 h3 Bf3!! 4 gf Qf2! 5 Qb7 Qg3!!- high class tip toeing, for sure.

Overall, an excellent puzzle: one for Aagaardian analysis, or for an article by Dvortesky.

Posted with Blogsy