Reverse Chess Engine?

Are there any chess programs that for any given board position 1) generate the list of possible board positions that could produce a given position after 1-ply (i.e. after one player plays one move) and 2) perform some standard engine analysis to rank those prior positions based on the centipawn loss of the move needed to produce the given position? It seems to me that this would allow one to take an interesting puzzle position that may or may not have been reached in a given game and by working backwards one might be able to create a relatively accurate series of moves that terminate with the given position and begin in known opening theory. One might even be able to tease out how an interesting middlegame trap position might be plausibly reached by playing certain openings.

https://www.chess.com/analysis

Snoozyman, how is posting the link to the chess.com analysis page an answer to my question?

Not sure what you mean, but you can analyze games and check computer for possible board positions. 

I want to know what board positions could produce a given position and which of those produce that position with the least centipawn loss this is in contrast to what the analysis page does which is to say that given a board position it finds what is the best next move (not the possible prior moves)

So what you want could be a first step on a generalisable algorithm to unscramble an egg. Even discover how life started, working backwards from the hugely puzzling position of humans who believe there is a god.

Wait...so you want to know if there's chess engine that can figure out how this:

....came from this:

By going reverse? That's a tough one, considering there are more chess games than the number of atoms in the universe...

Not really possible to many outcomes

Well I specifically said one-ply meaning one move by one player so I am not asking for it to reconstruct the whole game in one step, but presumably, it could be applied recursively. I am not sure why this is being met with such skepticism this seems to me to be symmetric to what chess engines currently do. If the images of board positions in your post were swapped around I could just pose the analogous question about current engines to you. You want it to go from the starting position to the ending position? And similarly, I could raise the point you made about there being more chess games than the number of atoms in the universe. Obviously, engines can't analyze every position past a few moves as it becomes unwieldy yet nevertheless, they are useful at analyzing potential futures moves. So why couldn't they be just as useful at analyzing prior moves? I don't see any obvious symmetry breaker between the two cases. Really I just want to take a few really hard puzzles that a friend found and try to construct plausible high-level games that feature those tricky positions and use them in a short-story that I am writing.

 "If the images of board positions in your post were swapped around I could just pose the analogous question about current engines to you. You want it to go from the starting position to the ending position? And similarly, I could raise the point you made about there being more chess games than the number of atoms in the universe."

There's a big difference between "best moves" and "possible moves".

Hi Fins5090, Interesting question.  Not sure this is the answer you are looking for, but endgame tablebases are made with a similar kind of logic.  As in they work backwards from a position to find the best possible moves.  Their computation relies on brute force rather than judging a position in centipawns, but I thought I'd mention it.  I read this a few days ago & found it pretty interesting.

https://en.wikipedia.org/wiki/Endgame_tablebase

Wait, it's difficult to read from the formatting.

What is a ply?

EDIT:  And I have to say, is this not called learning Chess?

I don't know anything about this retrograde analysis stuff, but I'd say you might be better off to take those puzzle positions and consult a large games database to see if the positions came from a real game. If they did, you'll have your high level games.

Exactly. There are websites that names each puzzle from which it originated. For example, the Bobby Fischer vs Mikhail Tal 1959 puzzles.

Yeah, I was just thinking more like installing Scid vs. PC and the Caissabase games database, then searching the database by position.