2024/03/08 DPA: "Clearing The Diagonal"

White to move.

Black has a problem at g7.  White has both Knights ready to spring into action.  And don't forget about the Bishop on b2, which could play a role.

1. Neg5 [threatening 2. Qxh7#] fxg5 opening up the long diagonal]  3. d5^ [discovered check] fails due to 3. ... Nxb2.

So can we preface our sequence?  1. Rxc4 and if 1. ... Rxc4, continue as above.

However, Black is up a Rook so is under no compulsion to recapture:  he could play 1. ... Rg8 instead, stopping the checkmate on g7 and potentially planning 2. ... Rg6 to chase the Queen away.

In order to win, White will have to keep generating threats of imminent checkmate.

Also, the light-squared Bishop might play a role if the Knights move and it can reach e4, which also threatens Qxh7#.

Nxf6 is bad because it allows ... Bxf6, which blocks the long diagonal.  I think Ng5 by either or both Knights will be in order.

And since puzzles can have only one solution for the solver, White has to figure out which Knight to move.

Aah, I think I see something:  1. Nfg5 fxg5  2. Nf6 Bxf6  3. Be4:  because Black's dark-squared Bishop is on f6, it prevents the f pawn from moving to f5 to block the Bishop attack on h7.

So I was initially wrong on 2 counts:  White's dark-squared Bishop did not play a role and Nf6 was not a bad move [actually, the 2nd point was predicated on the 1st so they're really just one error].

But I was correct that it mattered which Knight to move to g5:  if White had used the other Knight, he would not have been able to play 2. Nf6.

This puzzle was similar to a recent one where we had 2 Bishops on the f file blocking our Rook from hitting f1 and the solution involved moving the Bishops such that White had no time to defend.

.

In the interests of full disclosure, I solved a very similar puzzle back in November and I recognized this as I was contemplating the Knight moves. I didn't remember the solution exactly and the original lacked a dark-squared Bishop for White [which simplified things since there's no need to think about checkmate along the a1-h8 diagonal].

Capablanca - Nimzowitsch, Bad Kissingen, 1928 from Vukovic's *The Art of Attack in Chess* [hat tip to @JohanVA for finding it].

@channelingcapa: I hope your channel was working today.