Draw by repetition in FIDE chess rules

[crem]

FIDE has quite a long history of trying to define the repetition rule, they update it every time, and still fail to come up with something unambiguous.
I've just discovered that in 2023 they updated the rule again, this time rolling back to the previous version with some changes -- but it's still bad.

It's surprising that it's possible to fail to write clear rules for so long, for something as strict and deterministic as chess.

So, before (in 2018 version):

9.2.2
Positions are considered the same if and only if the same player has the move, pieces of the same kind and colour occupy the same squares and the possible moves of all the pieces of both players are the same. Thus positions are not the same if:

9.2.2.1 at the start of the sequence a pawn could have been captured en passant
9.2.2.2 a king had castling rights with a rook that has not been moved, but forfeited these after moving. The castling rights are lost only after the king or rook is moved.

Current (2023):

Positions are considered the same, if the same player has the move, pieces of the same kind and colour occupy the same squares, and the possible moves of all the pieces of both players are the same.
Positions are not the same if a pawn that could have been captured en passant can no longer be captured in this manner. When a king or a rook is forced to move, it will lose its castling rights, if any, only after it is moved.

Questions:

1. What does "possible moves of both players" mean? At any given moment, only one player has right to move -- so what "both" is doing here?
2. If a player has right to castle but cannot execute it (e.g. there is an own piece between a king and a rook), and then the same position repeats but the right to castle was lost by then -- it is a repetition? (after all, the set of "possible moves" is the same).
3. Some people argue to two points above that "possible moves" are not in the current position, but rather all future possible moves in the game. -- but then if we approach the 75-move forced draw horizon and therefore lose ability to do some future moves -- does it cancel the repetition?
4. If (during the first repetition) a pawn "can" capture en passant, but not really, because it would open a check to the king, is it a repetition or not?
5. In the last sentence of the rules, what if a king or a rook are not "forced" to move but just moved without being forced? (What does "forced" mean at all? Rules don't define or use this term elsewhere. If it's not because of a check but because all other pieces are blocked -- is it forced?)

I do understand what they wanted to say, but it's different from what they actually wrote.

[syzygy]

I do understand what they wanted to say, but it's different from what they actually wrote.

How would you write it?

[syzygy]

I do understand what they wanted to say, but it's different from what they actually wrote.

How would you write it?

Well, let me answer this.

Two positions occurring in a game at different moves are the same for the purpose of the repetition rule if:
- same side to move in both positions;
- the same pieces are on the same squares in both positions;
- each side has the same castling rights in both positions;
- capturing en passant is not possible in either position.

[Brunetti]

How would you write it?

The two FENs are identical

Alex

[syzygy]

How would you write it?

The two FENs are identical

Ignoring the move and ply numbers.

Unfortunately the original FEN specification does not deal optimally with en passant. A position after a double pawn push will have the en passant target square in the FEN, but if en passant capture is not possible or legal, the postiion is the same as the position corresponding to the same FEN but with the en passant target square removed.

But if you fix this (ep target square only if en passant capture is possible and legal), and you also require that the FEN "counts" all empty squares (I think most FEN readers accept "//" for "/8/", etc.), then it should work.

[Brunetti]

Unfortunately the original FEN specification does not deal optimally with en passant. A position after a double pawn push will have the en passant target square in the FEN, but if en passant capture is not possible or legal, the postiion is the same as the position corresponding to the same FEN but with the en passant target square removed.

It's impossible for the en passant capture to be impossible in one position and possible in another with the same FEN (excluding the move counters as you mentioned).

Alex

[crem]

Two positions occurring in a game at different moves are the same for the purpose of the repetition rule if:
- same side to move in both positions;
- the same pieces are on the same squares in both positions;
- each side has the same castling rights in both positions;
- capturing en passant is not possible in either position.

"Same pieces" should be replaced with "pieces of the same kind and colors", as if e.g. two rooks or two knights of the same color swap their positions, it's still a repetition.
Otherwise, it looks good. So it's not that hard to some up with clear rules. I would understand if FIDE had it ambigous in one edition of rules and then fixed in the next -- but they change it every time, seemingly trying to make it clearer -- but every time it's not any better than before.

Also it may be a question whether "illegal en passant" makes a position different (I believe the agreement is that it doesn't).

[syzygy]

Two positions occurring in a game at different moves are the same for the purpose of the repetition rule if:
- same side to move in both positions;
- the same pieces are on the same squares in both positions;
- each side has the same castling rights in both positions;
- capturing en passant is not possible in either position.

"Same pieces" should be replaced with "pieces of the same kind and colors", as if e.g. two rooks or two knights of the same color swap their positions, it's still a repetition.
Otherwise, it looks good. So it's not that hard to some up with clear rules. I would understand if FIDE had it ambigous in one edition of rules and then fixed in the next -- but they change it every time, seemingly trying to make it clearer -- but every time it's not any better than before.

Also it may be a question whether "illegal en passant" makes a position different (I believe the agreement is that it doesn't).

And maybe "en passant is not possible" should read "en passant is not possible or not legal".

I agree it is not hard to get it right. But this may be easier for chess programmers than for a FIDE lawyer

[crem]

Someone pointed to me that the rules which I mentioned as "current" are actually from the 2009 edition, and in the 2023 edition they are the same as in 2018.

They do seem better, but still has issues:
- Same "all moves of both players" clause.
- weird "at the start of the sequence" phrase -- what is the "sequence" there? We are comparing positions. There's no other occurrence of the word "sequence" in the rules.
- I'm not sure I understand they wanted to say with "a king had castling rights with a rook that has not been moved, but forfeited these after moving", particularly because the word "moving" seemingly syntactically only relate to a king, and not the rook.

[Brunetti]

- I'm not sure I understand they wanted to say with "a king had castling rights with a rook that has not been moved, but forfeited these after moving", particularly because the word "moving" seemingly syntactically only relate to a king, and not the rook.

In practice, it makes no difference because whichever piece is involved, the right has been lost, so the ambiguity in the semantic relationship is perfectly acceptable

Alex