Why is this game not a draw?

[bob]

It is not a 3-fold rep yet. The position after 12.Nxf6+ is NOT the same as those after 14.Nf6+ and 16.Nf6+:

r1bqkb1r/5p1p/p1np1N2/4p3/4P3/p4Q2/PPP2PPP/R3KB1R b KQkq - 0 12

versus

r1bqkb1r/5p1p/p1np1N2/4p3/4P3/p4Q2/PPP2PPP/R3KB1R b KQ - 8 16

Spot the difference! So Arena is right, and Xadreco's rep-detection is buggy.

I tried this on crafty, and you are correct. I thought it was "same moves possible in all 3 positions, and in the first, castling is not possible since you can't castle out of check. Apparently my memory failed me here, but not when I wrote the code in Crafty...

[Desperado]

[d]3r4/8/8/8/3p3K/8/R3P2k/8 w

9.2
The game is drawn upon a correct claim by the player having the move, when the same position, for at least the third time (not necessarily by a repetition of moves):

a. is about to appear, if he first writes his move on his scoresheet and declares to the arbiter his intention to make this move, or

b. has just appeared, and the player claiming the draw has the move. Positions as in (a) and (b) 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.

Hehe , i think HG is right. After the move e4 it is simply not possible
to capture en passent, so the repetition occurs.

Without checking the king, the repetition would be delayed because of
the oppertunity to capture en passent.

Funny , i like it...

[Sven]

I think it is, because it will also never have the opportunity afterwards. For e.p. the right is the opportunity. That makes the 'e.p. right' completely illusory. In fact you have no right to make an e.p. capture here, just like you would not have that right to make it if there was no Pawn on d4, or if the Pawn on d4 was pinned. It doesn't matter what the reason is that makes e.p. capture an illegal move.

For castling rights, however, there exists a right that is distinct from the opportunity, meaning that there exist positions where you cannot castle now, but might be able to do it later. (Like in the opening position.)

I fully agree. The current FIDE rules explicitly make this distinction between en passant and castling:

9.2 [...]
Positions as in (a) and (b) 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.

So the en passant example given by HGM above is definitely an example of a repetition.

Regarding the role of castling rights the interpretation of the rules goes like this:
Even though the first sentence quoted above says "... the possible moves ... are the same", the last sentence regarding castling rights only makes sense if those castling rights are included in the (somewhat imprecise) term "possible moves". This is also how the rules were always handled: if two positions differ only in the castling rights but not in the set of legal moves available at the moment then they are still different. E.g. the position after 1.Nf3 Nf6 from the starting position (let's call it "position A") is *not* repeated by moving 2.Rg1 Ng8 3.Rh1 Nf6 (let's call this "position B") since in A white has the kingside castling available but not in B. This means: castling rights are considered when deciding about a repetition. The sentence "When a king or a rook is forced to move, it will lose its castling rights, if any, only after it is moved." is a clarification of one special case and is fully consistent with the rest of the rules.

Sven

[Werner]

The Shredder GUI announces a 3-fold repetition after the move
16 Sf6 here,

Fritz GUI after 16. ....Ke8-e7

Werner

[Sven]

The Shredder GUI announces a 3-fold repetition after the move
16 Sf6 here

The question is, does the Shredder GUI announce a repetition based on the position after 16.Nf6+ (which would be wrong as stated above), or based on the possible move 16...Ke7 leading to a repetition?

Sven

[hgm]

I guess I really will have to adapt WinBoard's rep-check code. If the positions are the same (board-wise and castling-wise) but differ only in apparent e.p. opportunity, (i.e. a Pawn double-push ocurred on the previous move, and there is an opponent Pawn standing next t it), it should really test it the e.p. capture of that Pawn is legal. I guess the cleanest way to do it is when deciding to set the e.p. flags (after every double-push); then there also wouldn't appear any spurious e.p. rights in the FEN when you do copy-position. (Although I know this is not an offence.)

[Mike S.]

I like this topic. These are typical bugs even some top programs have had, in the 1990s (and beyond?...)

[D]r3k3/2Qp3R/1p6/1b2K3/4p3/8/5p2/8 w q - 0 1

This is a #4 which includes a semi-repetition but black loses his long castling right. I didn't check many current engines, but it looks like even some of the better ones will produce "suprising output", here.

[Event "Die Schwalbe"]
[Site "?"]
[Date "1929.??.??"]
[Round "?"]
[White "Zepler # in 4"]
[Black "[+1430.04e5e8]"]
[Result "1-0"]
[SetUp "1"]
[FEN "r3k3/2Qp3R/1p6/1b2K3/4p3/8/5p2/8 w q - 0 1"]
[PlyCount "5"]

1. Kd4 {threats 2.Qe5+ and 3.#} (1. Qd6 $2 O-O-O) 1... Ra4+ 2. Ke5 Ra8 3. Qd6
1-0

[Michel]

This is the official phrazing. It is still slightly ambiguous.

Positions as in (a) and (b) areconsidered 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.

[Carlos777]

Another one by the same guys. Same position at moves 15, 17 and 19, except that Mystery has both castle rights at move 15. Definetely a bug by Xadreco.

[Event "Div3s3"]
[Site "CARLOS-PC"]
[Date "2012.10.29"]
[Round "4"]
[White "Mystery 2.1"]
[Black "Xadreco 5.7.0"]
[Result "1-0"]
[BlackElo "1100"]
[ECO "A31"]
[Opening "English"]
[Time "20:35:35"]
[Variation "Symmetrical, Two Knights, 3...g6 4.Nc3"]
[WhiteElo "1100"]
[TimeControl "80/300:80/300:80/300"]
[Termination "time forfeit"]
[PlyCount "37"]
[WhiteType "program"]
[BlackType "program"]

1. Nf3 Nf6 2. c4 c5 3. d4 Nc6 {+1.30/4 2} 4. dxc5 {+0.26/6 4} Ne4 {-0.20/3
2} 5. Qd5 {+0.93/7 3} Nf6 {+0.60/4 3} 6. Qg5 {+1.05/6 3} Ne4 {+0.70/4 3} 7.
Qe3 {+0.98/7 3} f5 {+1.00/4 3} 8. Nc3 {+1.26/6 3} Nxc3 {+1.30/4 3} 9. bxc3
{+0.96/7 3} b6 {+1.09/4 3} 10. cxb6 {+1.16/7 3} axb6 {-0.01/5 3} 11. Nh4
{+1.66/7 3} Ra5 {+1.10/4 3} 12. Qf3 {+1.71/6 3} e6 {+1.59/4 3} 13. Qh5+
{+3.70/7 3} g6 {-0.41/5 3} 14. Nxg6 {+5.77/7 3} f4 {-0.92/4 3} 15. Ne5+
{+5.03/7 3} Ke7 {-7.49/5 3} 16. Qh4+ {+6.76/7 2} Ke8 {-0.52/6 3} 17. Qh5+
{+5.03/7 3} Ke7 {-7.49/5 3} 18. Qh4+ {+6.76/7 2} Ke8 {-0.52/6 3} 19. Qh5+
{+5.03/7 3} ... {Black forfeits on time} 1-0