Can you solve this helpmate?

Here's a retro which is likely cooked (please comment if you find one). Hope you like it

Any help in making this a standard helpmate would be greatly appreciated.

Isn't this a helpmate in 1.5, not 2.5?

Based on retrograde, I believe it is impossible for black to be able to castle since white has to capture a rook to get the a5 and a7 pawns there.

Even if white doesn't capture a rook, black doesn't appear to have any way to make any waiting moves with anything but the king and rooks. For example this position:

And black must move the king or a rook to let white capture on a7!

(1) it's not "after Keym" since that only affects who starts and not the relation between castling and e.p. Standard "A.P." will do for this one

(2) I do not understand the instructions. "Contingent on" is a logical expression which has no clear meaning in chess. Do you demand that white captures pawn b5 before checkmating? If so you would write "H#2.5 while capturing pawn b5 in the process". Which is a contrived stipulation form. It evades the duty of the composer to disable all dual solutions with technical means. Like I could say "#3 while the white knight does not visit square c3 and his queen stays at home". Very artificial. But perhaps you meant something different?

Is this the solution?

White could have captured pawns on d3 and a5, bishops on c3 and c4, knights on b4 and b6 and a queen on a7.

Backing it up a couple of moves the White pawn could have been on b6 instead of a7, the black pawn could have been on c6 instead of b5, and put the white queen on b5 with the black queen on a7.  It is a position reachable with either player on the move, black still being able to castle and en passant not available.

Thanks for the comments!

@BishopxH7, the instructions were a bit unclear. I meant the white had to capture the b5 pawn in the solution of the helpmate. And although the position is legal, I did fail the retro part of the puzzle.

@Arisktotle, thanks for the correction! Yes, I meant "while capturing b5 in the process). I agree with you when you say not to use technical means to disable solutions, but I doubt it is possible in this case - the rook on a6 is needed for the retro component, and this means white can play things like "0-0 Rd8 a8=Q Kd7 Qc6#" or "Rb6 0-0-0 Rb7 b3 a8=Q". 

Anyways, the solution is cooked - I did not see that Pa7 didn't have to promote. Thanks for the comments!

Actually, I meant kind of the opposite A composer would be expected to tackle obstacles underway by finding clever technical chess devices in the real play and in the tries and not by complicating the instructions of the stipulation. In concreto, the whole section about the pawn on b5 should not be there. I believe you when you say that it is then hard or impossible. If so, then that is simply what it is and your idea may not be realizable. Do you have any idea how many times every composer has thought "if only I had an extra square h9 or an extra dark squared bishop or ....)". It's a matter of discipline not to give in to such feelings and continue to walk the thin blue line..

The real issue is somewhere else. It is tremendously hard to create decent A.P problems - just as it is very hard (but easier) to create an original retro-analytical challenge of the type "release the position". Not many of those are made anymore as most interesting concepts have been shown.

That does not terminate the field. Instead of adding ad hoc conditions you can define complete environments which are way more fruitful for AP-logic and even traditional retro-analysis. But it requires that you step away a bit from the conventional chess rules. For instance, use Chess960 which has great retro-potential due to the variations in the PAS (starting position of the game). I will post an example later of an A.P. problem with a slightly different logic. Perhaps you will see where you gotta go to find unexplored planets in the solar chess system.

Well, here is an example of A.P.  It is of the standard type "e.p. versus castling". However one of the basic conventions has been turned upside down, the e.p. convention. In this challenge the e.p. permission is dominant which means you are permitted to play e.p. whenever there is at least one proof game for it, comparable to the standard castling permission. 

The stipulation is R#1 is very common amongst problemists but probably not for you. It means  reflexmate in 1 move which is the same as selfmate in 1 move with the extra condition that both sides must mate whenever possible. As in selfmate, white starts and forces black to checkmate him. Obviously, black must cooperate on his last move because the reflex rule demands it.

Try to find it for a while which is possible when all the rules are understood. But, in the end, I can prove the solution to you. And there is a decent try as well!

@Arisktotle #10 Ah, I see. Don't add weird tasks onto the stipulation when you can help it. Rather, change the type of chess environment for better compositions! 

#11 is very beautiful, and I think I have the solution (in white text, below)

White wants to play Qe5, where black must then reply with 0-0#. However, white is also obliged to give mate when he can, and since there is a possible PG where the last move was b7-b5, white cannot play Qe5 but must play axb6 e.p.

Lets work back from the b7-b5 PG. The sequence of moves must have been c6-c7+ b7-b5. However, looking at the position before c6-c7, we see black's only last moves could have come from the king or rook! If black has castling rights, b7-b5 does not have a possible proof game (rather, the sequence of moves would be e.g. Bb5-a4 b7-b6 c7-c6+ b6-b5).

Because of the AP in the stipulation, all we need is to get black to castle, and then we can prove e.p. is not possible. And this is the solution: Qe5! (anyway). Black is forced to castle because of the reflex stipulation, thus disproving that e.p. was possible.

Also, one of my earlier compositions, a seriesmate in 3, is also cooked.

Instead of the intended h7-pawn goes to g4, g7 pawn goes to g5, and e7 pawn promotes (after capturing a piece) and aids in getting white's pawn to h7, h7 can stay where it is, g7 goes to g4, and e7 captures twice to get to g5. Both take two captures, so black's last move can also be f6xg5!

That's amazing! I didn't expect you to get the idea so quickly! Everything you write is true but there is one "little detail" left. If white starts with 1.Qc3!? then black can checkmate with Bxc3# (0-0 is not mate). However that does not disprove the e.p. permission which was therefore mandatory instead of 1.Qc3!? I am very happy with this thematic try!

I don't exactly get the problem with your earlier composition because I see no seriesmove anywhere in the analysis column. What's the idea?

I missed the try! Very nice!

https://www.chess.com/forum/view/more-puzzles/can-you-solve-this-seriesmate

Comment 20, intended solution was hxg6 e.p. 0-0 g7, but black's last move doesn't have to be g5

Now I am even more confused. There are 3 different but somewhat similar diagrams (1 in the link) with 2 different stipulalions. What are you trying to show?

Btw, Reflex chess and Capture chess are powerful chess variants in combination with retro-creations though largely unexplored yet. The reason is that they contain rules to force/reduce options of play. Standard chess is a bit too free by nature to open up a really big playing field for retro-analysis and retro-logics.

Here is an example based on Fool's mate in combination with Reflexchess. It's a simple SPG (shortest proof game) challenge I just produced. Verify that this is dualed without reflex rule.

@Arisktotle I am trying to show:

Your PG is simple but sweet. I got stuck for a bit on why not the early g4, but the answer's in the comment

Capture chess seems like a fun area of exploration! Indeed, the OP could be tweaked a bit to make a decent Capture Chess composition (with the e.p. capture first move being one of the only captures possible)

Keep in mind that "reflex chess' and "capture chess" are game rules which apply to the past (the proof game) a much as the future. Producing proof games is mostly easy with reflex chess but  hard in capture chess.

Another way is to interpret "reflex chess" and "capture chess" as stipulation types. In that case they only start operation in the diagram. You already know this for the seriesmovers which generally assume that the preceding game is played with individual piece moves. Since the combination of "reflex" and "capture" with retro is relatively new there is no established practice yet in the field. Often as a composer in the retro field you play a part in defining and popularizing new types which keep coming hard and fast. Btw , in 2012 I won a 1st prize with a retro with the reflex rule!