Presented at my chess club - Mate in 5

White to move.

It's a beautiful problem - had you posted it correctly. The white pawn is really a rook and the position is upside down. Detail matters!

You were right about the board... was upside down, I fixed it.  But the white pawn is correct. The king is in way of its promotion - part of the problem. 

There's a convention in composed problems that diagrams are shown from White's point of view, which makes sense since it's normally White who mates in X moves.

Re the WP on e7, that's in the original problem, but Arisktotle (who's an expert problem composer in case you don't realise) has found a better setting of the problem by replacing the WP with a WR. Having the WR doesn't affect the main variation of the solution (other than that a WK move threatens 2.Re8+ rather than 2.e8=Q+), but in a secondary variation, the weakness of a dual is removed. Here is the solution:

The try 1.Nd5? threatens both 2.Nxc6 and 2.Nd7, but 1...Bxe5! refutes. So White aims to disable ...Bxe5 with a preparatory manoeuvre. 1.Kf8! (threat: 2.e8=Q+) Rf1+ 2.Bf4 Rxf4+ 3.Ke8 (threat: 4.Nd7) Rd4. We are back to the diagram position except that the BR has been decoyed to d4; now 4.Nd5 works as the BR has cut off the BB on a1; 4...Rxd5 5.Nxc6 or 4...Bxd5 5.Nd7.

The secondary variation is 3...Rf8+ which allows a dual, i.e. alternative white moves that equally work, namely 4.Kxf8 and 4.exf8=Q. If in the diagram, the WP is replaced by a WR, there's no promotion option and the unique continuation is 4.Kxf8. Although there are still some 5th move mating duals, this is still an improvement. Nice job, Arisktotle!

@Rocky64: Thnx! it's the 1e time I made an unintentional composition! I must have made a mistake in checking the one with the pawn (I thought it was wrong) otherwise I wouldn't have started on the rook attempt! I was happy with that because the back row felt like "a rook mate", not because it improved over the pawn version.

I suppose the composer might have considered the rook on e7 but finally chose for minimizing the material. It is one of those borderline choices. I agree mine has a bit more content.

@ScotchYa: You changed the coordinates but it is still confusing as the board remains upside down. Problemists only use the orientation with the white side down because there are no coordinates in standard problem diagrams and problem positions are often so weird that you can't know which way either side plays - different from game diagrams. Well, it's your prerogative as the OP to choose your presentation, but considering that your diagram is taken from the world of "professional" compositions, it is worth to take its conventions into consideration.

How long til you post the answer?

Thanks, good to know for future reference on setting the correct view.  

OK, something to learn here for you about "professional" compositions - I presume. There is only one solution to a composition and you must find it because it is also the only line that shows the idea of the composition. Without it, the whole puzzle crumbles - however fancy the alternatives look.

Of course the one solution may have variations because black can have more than 1 defense and often these variations show variations to the idea (theme) and logical derivatives.

Finding the solution and seeing that the solution is correct in itself means nothing. A composition should be understood to the point that you see the idea and logical structure of the solution which includes why the alternatives fail. In chess compositions, ideas/themes play a similar role to plans in a chess game. Designing the right plan is what separates good players from bad players. Recognizing the right idea of a composition is what separates good solvers from bad solvers.

So.. with all that said... do you have the 5 move mate solved?

The best white can do is seven moves; there is no forced mate in five moves. There are ways to checkmate black in as few as two moves if black plays poorly [1.Nxd1 Bh3?? 2.Nxc6#], but puzzles are supposed to be solved under the supposition that the opponent plays the best moves possible to prolong a loss as far as possible.

As others have mentioned, standard puzzle design is from the perspective of the side you play. If a puzzle is made with the intention of being misleading, then it distracts from learning the tactics involved. You need to explain to the person who presented the puzzle to your club that puzzles are meant to be instructive, not haha-gotcha's.

Here's your position in puzzle form.

Notice that I stopped at move five, not because you wanted a five-move solution, but because there are multiple ways to win in the same number of moves after this point. When black continues with [5...Nf7], white has the option of promoting to either a queen or a rook and the win will still happen in the same number of moves. It is only after these first five moves that the correct play only had one possible line. If you program the puzzle here to accept only the line where you promote to queen, then someone who attempts to solve it by playing [6.e8=R+] will be given an "Incorrect" notification despite their choice being an equally valid solution.

Good luck with your chess future!

I always attempt to solve puzzles and mostly succeed. However, because your original presentation was flawed I worked on discovering what you did wrong. In my role as detective I first feed my attempted settings to an engine. I would waste a lot of time if I were to solve each change from zero only to find out that the setting still isn't right. So the engine confirmed that I was on the right track and then I studied the solution to see that it looked like a proper composition. And it did!

As a puzzle poster you ought to check every puzzle you post by engine and study the solution as I did. That the engine says "mate in 5" is not sufficient because there may be duals (more solutions). It is very frustrating for solvers in general to invest their brainpower in analyzing a puzzle which is't right. I personally don't mind so much because "detective" is my second nature  

Unless, of course, there is ......

You are right in that the board orientation should not be used to deceive the solver. Best to use standard orientation always. But I don't think the OP tried to deceive us. He just didn't know the convention.

In general, a promotion dual like Q/R is not considered fatal in composition chess though some points may be deducted depending on the total content of the composition. It becomes a problem for a solver using the chess.com puzzle interface because it only accepts 1 move as correct. It is best when we agree that we always promote to queen when it works. That resolves the issue practically.

This is wrong... I did end up putting in the engine and it is a mate in 5. 

The 5 moves.

Btw, the black pawn on c2 is unnecessary. It adds nothing to the solution and the problem is still correct without it. I assume it wasn't there in the composer's original.

Having said that, I just discovered something interesting. In my version with the rook on e7 rather than the pawn, there must be a black pawn on c2 to refute a dual solution. Which implies that - if there is a pawn on c2 in the original - there will also be a rook on e7! Still assuming it is an optimal composition of course.

It pays to be a detective!

@Arisktotle According to a database, this problem was composed by Hans Haden in 1969. There is a WP on e7 and a BP on c2.

I've computer-tested both the original and your version with the WRe7, and here's what I found.

(1) The c2-P seems to be there to stop an unwanted threat in the main line. After 1.Kf8!, the only threat should be 2.e8=Q+ but without the black pawn, 2.Nxd1 would be threatened too, making things slightly messier in some lines.

(2) As you indicated, in the WRe7 version, the c2-P is also needed to stop the cook, 1.Rd7. With the c2-P in place, 1.Rd7? becomes a try that's defeated solely by 1...Nf7! This try generates a ridiculous amount of variations, though they don't seem that interesting.

(3) In the original, after 1.Kf8!, if 1...Rd8+ then of course 2.exd8=Q, but in the WR version, we get the nicer 2.Re8 Rxe8+ 3.Kxe8.