As far as the issue goes regarding castling with a phantom rook: castling is defined as moving the never-before-moved king two squares toward the never-before-moved rook on the same rank (the rank qualifier was added to prevent castling with a pawn promoted on the e-file to a since-unmoved rook) and moving the rook to the square the king crossed over. If there is no rook to move then no castling can be done (no direction for the king to go and no rook transfer at the end).
As far as temporary versus permanent loss of castling rights go, I feel that permanent loss makes a different position than a temporary loss.
Better to change a rule to be clear and consistent, than to need a TD Tip. Those should be left for applications.
Why does castling have a “right”, when all of the other potential moves don’t?
There was a very, very old rule, that if you touched a piece, and couldn’t move it (probably due to a pin?), you had to move your king. “Rules of Chess” is a developing document.
This quoted opinion is exactly why I started this subject. In comes down to consideration of the available moves at the time of the claim or the potential available moves in the future. Without the insertion of the word “permanent” into the rule, two interpretations are possible.
I don’t know if it’s going to make much of a difference as this situation has happened to me only once in 34 years of directing. I estimate that this includes close to 2000 tournaments and 200,000 games.
Now that I’ve thought about it more, I’m going to make a case for the other side:
It could be argued that the issue is not just what moves can be made (“in the given position, then-and-there for next move”), but also what considerations there are for making each of those moves. At move 17, one of the reasons White has for choosing Qh8+ (rather than, say, Qxh6) is that he wants to keep Black from castling. But this reason does not exist for moves 19 and 21. So is the board position after move 16 really the same as it is after moves 18 and 20?
I interpret 14C as being indifferent to whether the right to castle has been lost temporarily or permanently. Either castling is included in the collection of legal moves at each occurrence of the position, or it is not, for whatever reason. More specifically, I interpret 14C as: “Each time the position in question occurs, write down a list of all the legal moves that could then be made, including the move that caused the position to occur. In all three instances, those sets of moves must be identical.” It does not matter why castling is prevented (if it is), and it certainly doesn’t matter why the player might want to play a certain move.
However, here is what I think is an interesting test of “the possible moves of all the pieces.” Consider the following hypothetical game: 1. g3 g6 2. Bg2 Bg7 3. Nf3 Nf6 4. Ng1 Ng8 5. Nf3 Nf6 6. Ng1 Ng8 7. Kf1 Nf6 8. Ke1 Ng8 9. Nf3. If Black now writes Nf6 on the scoresheet for his ninth move and claims a draw by triple occurrence (not “triple repetition”!) of position, how would you rule?
The claim at move 9 is invalid because in the prior occurrences of the position White had the right to castle. That one is explicitly covered in the rules. Note, however, that a claim at move 6 would have been upheld.
Ah, but there’s the rub. White having lost the right to castle does not affect the set of legal moves when Black is on the move. If I apply my “list all the legal moves in the given position” theory, with Black on the move, the set of legal moves is unaffected by White having permanently lost the right to castle.
Now, had White attempted to claim a draw by writing 9. Nf3 on his scoresheet, I would absolutely have denied the claim, exactly because the set of legal moves is different in that case.
14C2 says " … the position is considered the same if pieces of the same kind and color occupy the same squares and if the possible moves of all the pieces are the same, including the right to castle and to capture en passant." Good try, though.
According to the rules, the right to castle cannot be lost temporarily. It is only lost by moving king or rook. 14C is explicit about the right to castle. Temporary prevention of castling (i.e. temporary loss of the ability to castle) is covered by the possible moves wording.
I may not be expressing myself clearly. I understand the reference to the right to castle and to capturing en passant in 14C. (I don’t understand the reference to 14C2, How to claim. I’m finding the quoted text in 14C, at the top of page 43 in the fifth edition of the rule book.)
Let me be more blunt: When Black claims the draw after writing 9. … Nf6 on his scoresheet, I claim that there are no possible moves for any White piece, since Black is on the move. Of course, it would make absolutely no sense for Black on his ninth move to pick up (say) the white knight on f3 and place it on g1. So, given that White’s castling O-O never appeared in the set of possible moves when Black was on the move, why does this affect Black’s draw claim?
In other words, the possible moves of all the pieces depends on which player is on the move. When Black claims the draw, I assert that the set of possible moves of all the white pieces is the empty set.
That’s not the point. “Possible moves” in this context means “legal moves now or at any time in the future.” The rule explicitly states that loss of the right to castle by either player means that the positions are not identical. This is why I remain uncertain about Harold’s example.
Steve, you are being much too deep for me to understand. I realize you’re illustrating the dependency of en passant capture on previous moves. That tells me that it is insufficient merely to look at the pieces on the board at any given moment in order to determine the set of legal moves. But I don’t understand how this relates to my contention that when one side has the move, the set of legal moves for the pieces of the opposite color is empty.
14C requires the same player to have the move, but apart from that I believe possible moves for all pieces is intended to be applied literally, i.e. you are not looking at the legal moves by the player who has the move, but at all possible moves for every single piece. FIDE is a little more explicit and mentions “of both players.” So the right to castle for the player who does not have the move is relevant.
If that were really the correct reading of the rule, then in any position (not just a constructed one) the loss of the castling privilege by the opponent of the claiming player from one repetition to another would not invalidate the claim. Since that is not the case (this has come up several times and both USCF and FIDE are quite clear on it), I don’t see how your reading can stand up.
The only case I can think of is where, say White, can repeat moves whenever he wants. After doing this twice he does something else and Black moves a pawn giving White a chance to take it e.p.Instead of doing that he repeats moves for the draw. In this case the position after Black’s pawn move isn’t the same as the position two moves later.
For castling, a player gives check with the rook on b1 or b8. After it returns to the corner it is no longer available for castling.
OTOH, the player responding to the check doesn’t have the option as others have pointed out.
I think the en passant case is different from my castling case. I hope we can all agree that in Steve’s composed example that en passant was legal and possible at the first time that position was reached (move 3), but wouldn’t be at any future point in the same game.
In the position I ruled on in 1984, castling was not legal at moves 16, 18, and 20. The difference is that it was a potential future move at move 16.
Ken’s point about not being legal when it’s not that player’s turn is something I hadn’t considered. If it is correct to ignore the potential legal moves of the player not on move in the claimed position, then I feel that I ruled correctly 26 years ago. On the other hand, if potential legal moves of the player not on move are to be considered, then I got it wrong.
It would be nice if the authors of the rule book contributed to this discussion with a clarification of the rule. On the other hand, it may be something they never considered until I brought it up. They may not yet have an answer. The fact that I started this subject a few days ago and no one has yet stated with 100% certainty what ruling is correct, indicates to me that this issue needs to be clarified.