Why does Houdini 1.5a fail on this position?

[Ajedrecista]

Hello:

I have to correct you.
This is a correct study! It's a study and not a problem so the longest line has nothing to do with its solution. White's play is dual-free except some time-wasting moves, so in Nunn's notation all White moves deserve an '!'.

Agreed.
In fact Black's longest defense might be the rather uninteresting 1.f7 Rg8 2.fxg8Q Kxg8 3.Ne7+ Kf7 4.Nxc8 followed by a mate in 31.
This means the starting position appears to be mate in 35.

Robert

@Arpad: I did not know that there was any difference between study and problem. Thanks for explaining it: learn is always useful! I hope I understand the idea (please correct me if I am wrong): focusing on key moves such as Ba3!, Kc1!, Kd2!, Kh6! or Ne7! (for example), despite black plays some moves (such as Rg4+) that lead to shorter mates. It is a pleasure that an study (not trivial for me) of sixty years ago is correct.

I thought Nunn's notation puts '!' when there is an only move that leads to a win in a winning position or when there is an only move that leads to a draw in a drawn position. I copy from English Wikipedia:

! – the only move which maintains the current evaluation of the position: if the position is theoretically drawn, this is the only move which does not lose; if the position is theoretically won, this is the only move which secures the win. An "!" is used no matter how trivial the move in question; the only exception is if it is the only legal move.

I did not add '!' in all the white moves because of this.

I have seen some of your studies and they seem very elaborated! Congratulations.

@Robert: This line is explained in the video I linked. I wanted to say that starting with 1. f7 Ra6+ 2. Ba3 Rxa3+, then the longest mate was in 18 moves, so with the starting position given it was mate in 20; anyway, I did not want to assert that the longest mate was in 20 moves, only that it was mate in 20 with the line I posted (starting in such way). In fact, this K + B + N vs. K is the longest mate.

Congratulations for the hard work in Houdini engine.

Regards from Spain.

Ajedrecista.

[Arpad Rusz]

The Wikipedia is wrong.

John Nunn - Secrets of Rook Endings :
"The exclamation mark ! after a move means that this move is the only one not to change the result of the position... The use of the word 'only' needs to be clarified. I consider a (white) winning move to be unique if any other move preserving the win allows a defence by which Black can force White to return to the current position in order to win. In other words suppose that White has a winning move, but he can also repeat the position before embarking on the winning continuation. In this case the winning move is still marked unique, even though the repetition also retains the win."

[Mike S.]

On the other hand, H15a finds a rare mate with two knights against a lone king in the following position, where some (old) engines chose Kxh6 with a 0.00 eval:

[d]7k/8/7p/4n2K/5n2/8/8/8 w - - 0 99

Houdini 1.5a:
---------------------------------------------------------------------------
3/8 00:00 187 0 -M1 99.Kxh6 Nf7+
3/8 00:00 187 0 -6.84 99.Kh4 Kg7 100.Kg3

(that's from multi-pv(3) but there are only two legal moves)