Illegal Position Contest!

Deleted message, click here to view

learn and watch 8/1p1ppppp/8/8/8/PkPpP3/1PpPP1P1/RqKB4 w - - 0 1

A much more difficult position (compared to my previous one).

Black's a-pawn captured thrice to reach d2. Pf7 captured on e6. This covers all missing white units. White promoted the 3 Kingside pawns, two on f8 and one on g8, capturing N and B on the way.

We can't retract d6 until Bc8 is back, and can't retract g6 until Bf8 is back.

Retract -1.Rd8 h4 -2.Rf8 h5 -3.f7=P h6 -4.f6 h7 -5.fxg5=N Nd5 -6.g4 Nb4 -7.Nc5 Na6 -8.g3 Rb4 -9.Nb3 Rh4 but white cannot retract.

If instead -8...Ra3 -9.Nb3 now Black cannot retract.

Position, as expected, is illegal.

Guys can I ask how to you come up with these so random yet so hard positions? They're almost impossible to figure out let alone create.

I've found usually I just mess around on a board trying ideas. Then I try and figure out how to make the position unretractable. And there's lots of ways to do that: pieces blocking each other, pieces getting trapped, etc.

Yes but alot of the explanations involve tracing back to the beginning of the game, such as contradictions involving how certain pieces could have gotten out from behind each other. Seems way too complicated to realize by just setting up weird cramped positions on the board.

Great work with the proof (and sorry for the delayed response).

The reason I thought that this might be tricky was that there were many ways to "un-pin" bRb3: it may have been by wRd7, or by the wR/N captured on e6 instead of wNa6. The try you showed was the most involved, but I thought it might have been a nice idea to include the other tries as well.

I usually start with a main idea/structure that I want to implement on the board. For example, I had this pawn structure which featured a narrow bishop passage in one of my previous positions (#8765), and the idea was to prevent the bishops from swapping places. After this, I fill the board with pieces to limit the number of captures for each side, tightening the inventory.

In some of my positions, I try to make one/both sides have limited tempo - in this case, I trap the other pieces on the board to prevent them from providing tempo moves. Here, the closest tries to legality typically require certain "events" to occur when retracting moves (e.g. un-capturing a certain piece, retracting a pawn back a certain distance).

For my other positions, the illegality is more focused on certain pieces and pawn structures. This often involves some moves that must precede others (e.g. b3 or d3 must be played to free wBc1), creating a chain of moves one before another. The illegality would then be caused when the chain circles back on itself, creating the paradox of "which came first".

You've also created some tricky positions where the solver has to prove a required castling would not be legal. I wonder if the same thing could be done for an en passant capture. I don't recall seeing one like that.

Or make those two themes the contradiction itself that makes it illegal. Maybe a position where en passant could be legal and castling could be legal but not both at the same time, if this were somehow possible.

Well, this is what I came up with. Either black can capture cxd3 e.p. or white still has their queenside castling right, but not both.

I've seen a number of published studies based on the interaction between en passant and castling, but of course in those the positions were all legal.

Here is a rather silly problem I came up with. When a game reached the position below, a spectator reported that White had just moved a pawn and said "Checkmate". Black replied "That's not mate", made a move, and said "Checkmate".

If the spectator's report was accurate, can you determine which player made an illegal move?

Black just played 0...f4xe3 ep, preceded by -1.e2-e4+

With pawns on e2 and g2, and all 8 white pawns on the board, white must have illegally jumped his bishop over one of the pawns en route to b5.

I totally missed the concept of the bishop getting out, was looking at the pawn doubling for a few seconds but yeah f and h pawns could have easily reached their current squares so that the e2 pawn was at home, but yeah then the b5 bishop couldn't have gotten out.

Either player might have made the illegal move. If the white pawn reached e4 from e3, Black's ep capture was illegal.

yes