chess online

To create a castling scenario, you must start out with a full chess setup. Then remove all the pieces you do not need and reposition the pieces you do.

yadasmpati: The point is that White cannot legally castle in the position, because his king had to move earlier to permit the rook from h1 to get to b4. Whtie couldn't have promoted to a rook because he still has all eight pawns. (I wish castling was legal; it would be a prettier solution!)

In competitions for problem composers, there are certain conventions for castling and en passant. First of all, positions have to be possible using the regular rules of chess, meaning that a legal game could be constructed leading to the position on the board. Castling is assumed to be possible unless there is proof it is illegal, as in the problem I just made up. En passant captures are assumed to be possible ONLY if it can be proven that Black's last move was pushing the pawn in question two squares.

Here is a simple example, composed by F. Amelung in 1897. White mates in two.



Black's last move had to be ...g7-g5 (the pawn couldn't have come from g6, where it would have been checking White's king). So White is allowed to play 1.hxg6 Kh5 2.Rxh7#

This is known as "retrograde analysis". Some find it fascinating, but many object to it as artificial and useless. Both points of view are right. I think the majority of users here prefer more gamelike positions and strategies, and don't want to bother with these subtleties. Also, the problem software doesn't handle these retro problems well. That doesn't mean we shouldn't have this kind of problem, just for variety's sake if nothing else. But if you're really interested, there are other sites around the web that you should visit.

As phonybenoni nicely explains, puzzle #74962 is actually illegal, because it has a position that can never be reached without the king having moved (otherwise how did the rook get on b5?). Thanks for creating it though, because it shows that the puzzle creation software has a functional bug in it!

And phonybenoni also gives a nice example of a puzzle in which the first move is an en-passant move. I also understand now why the solution to the problem he gave earlier (the color-reverse of puzzle #74962) can never be 1. 0-0-0+ because the king must have moved to reach this position. But with FEN-notation the castling-option can be forced.

Thank you both for your efforts!

check puzzle #59265

Do you know how this ep function was entered into this puzzle?

Well, i was about to try it out later. But you might save me the time and effort. No i do not know  

This one's pretty cool.

puzzle #27402

The comments with puzzle #58981 give a clue as to how the en-passant move is accomplished on the first move. I'll endeavor to pursue it further tomorrow when I have more puzzles available to create.

This part " - - w" may also explain why GameKnot's analysis board is confused and show 1. ... f7 instead of 1. f7. It is not the right solution because black plays 1. ... g4!

How do you find these puzzles? I have to rake through past puzzles to locate what I'm looking for. Is there a search engine that I am not aware of?