FRC format question and wrong PGN from Mainz

[Volker Pittlik]

After converting some PGNs from Mainz using the latest version of Franz Hubers's conversation tool I got

[Event "Clerical Medical Chess960 World Championships"]
[Site "Chess Classic Mainz"]
[Date "2006/8/15"]
[Round "1"]
[White "Kosteniuk"]
[Black "Paehtz"]
[Result "1-0"]
[SetUp "1"]
[FEN "rkqnnrbb/pppppppp/8/8/8/8/PPPPPPPP/RKQNNRBB w FAfa - 0 1"]

1. e4 e5 2. Nd3 g6 3. f4 exf4 4. Nxf4 Nd6 5. Bc5 f5 6. Nc3 Bxc3 7. dxc3 fxe4 8.
Qd1 Re8 9. Bxd6 cxd6 10. Qxd6+ Qc7 11. O-O-O Qxd6 12. Rxd6 Kc7 13. O-O-O Re7
14. g3 Nc6 15. Bg2 Rd8 16. R6d2 d6 17. Re1 Rde8 18. Nd5+ Bxd5 19. Rxd5 Re6 20.
b3 Ne7 21. Rd4 e3 22. Rc4+ Kb8 23. Rb4 b6 24. a4 Kc7 25. a5 d5 26. axb6+ axb6
27. Rf4 h5 28. c4 dxc4 29. Rxc4+ Kb8 30. Kb2 Rc8 31. Rxc8+ Nxc8 32. Kc3 Nd6 33.
Bd5 Re7 34. Kd3 Nf5 35. c4 Nh6 36. Rxe3 Rxe3+ 37. Kxe3 Ng4+ 38. Kf4 Nxh2 39.
Be4 g5+ 40. Kxg5 Nf1 41. Kh4 Nd2 42. Bc2 Nf1 43. Bd3 Nd2 44. b4 Nf3+ 45. Kxh5
Ne5 46. Be2 Kc7 47. g4 Nc6 48. g5 Nxb4 49. g6 Nc6 50. g7 Ne7 51. Kh6 Ng8+ 52.
Kh7 Nf6+ 53. Kh8 1-0

as output.

Besides the fact Franz's tool (http://www.geocities.com/fhub55/chess/) is running fine
under wine and seems to be great at removing useless headers
my first question is if castling shouldn't be written as KxR if the
option /=s is given? Or is that only used for the communication
between the engine and the GUI?


However, Tord's FRC enabled Xboard can read the PGN above but it
complains about the castling (that was the reason for my first
question).


Then I took a look at the PGN again. Especially white's 11th and
13th move is very interesting. And that is what Xboard is
complaining about.

Volker

[Volker Pittlik]

... Especially white's 11th and
13th move is very interesting....

It's even on their online replay site (http://chesstigers.de/partien/ccm6/01_wc_dsj/01_20060815/live.htm) and their software allows to replay that. If that moves really happen then I didn't understand FRC, PGN is wrong or an illegal move have been played.

Volker

[Guenther Simon]

... Especially white's 11th and
13th move is very interesting....

It's even on their online replay site (http://chesstigers.de/partien/ccm6/01_wc_dsj/01_20060815/live.htm) and their software allows to replay that. If that moves really happen then I didn't understand FRC, PGN is wrong or an illegal move have been played.

Volker

Oops she castled two times in the same game?

Guenther

[Volker Pittlik]

...
Oops she castled two times in the same game?

Maybe a new FRC variant?

Volker

[Roger Brown]

Oops she castled two times in the same game?

Guenther

Hello Guenther, Volker,

What seems to be the problem here? It has been done before and in normal chess at that!

Later.


http://www.xs4all.nl/~timkr/records/rec ... 0castlings

[Guenther Simon]

Oops she castled two times in the same game?

Guenther

Hello Guenther, Volker,

What seems to be the problem here? It has been done before and in normal chess at that!

Later.


http://www.xs4all.nl/~timkr/records/rec ... 0castlings

Hi Roger,

I see a big difference to your example.
In the game at Krabbes site there were 46! plies between
the castlings of the same side. Here there are just 4 mere
plies between them. I am quite sure top WGM/WIM players
have a better memory than this
It must be a crippled pgn, which is of course no good advertising
for FRC, as it comes from the main FRC pages in Germany.

Best regards,
Guenther

[Volker Pittlik]

...
It must be a crippled pgn,...

The most likely explanation. Strange enough their software on the website allows to replay this long castle to the king side.

Volker

[Guenther Simon]

...
It must be a crippled pgn,...

The most likely explanation. Strange enough their software on the website allows to replay this long castle to the king side.

Volker

I looked at the game itself now at their site.
It is clear that their software wrote that O-O-O instead
of the real Rgd1! Big bug;-)

Guenther

[F.Huber]

Hi Volker,

Besides the fact Franz's tool (http://www.geocities.com/fhub55/chess/) is running fine
under wine and seems to be great at removing useless headers
my first question is if castling shouldn't be written as KxR if the
option /=s is given?

well, there are many different ways to write a castling move in FRC and there?s definitely no standard yet - some use KxR (but which R if both castlings are possible?), others use O-O and O-O-O (which I think is the more logical way).
So I?ve decided to not change the castling notation within the movelist at all - my tool does only normalize the FEN-tag in the PGN-header to the wanted format (Shredder/compatible/X-FEN).

Regards,
Franz.

[Volker Pittlik]

... some use KxR (but which R if both castlings are possible?), others use O-O and O-O-O (which I think is the more logical way)...

That makes perfect sense to me. For humans it is readable and needs no adaptation. For the communication between GUI and engine f1xg1 (if the king is on f1) for example would eliminate the possible ambiguity.

Volker

[Volker Pittlik]

...
It is clear that their software wrote that O-O-O instead
of the real Rgd1! Big bug;-)...

Meanwhile I have been informed by somone reliable that the software indeed has a bug and is writing incorrect PGN. They hope to get an updated version soon (tomorrow).

Volker