Annotation formats for CEGT-CCRL-TCEC

Discussion of chess software programming and technical issues.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Adam Hair
Posts: 3201
Joined: Wed May 06, 2009 8:31 pm
Location: Fuquay-Varina, North Carolina

Re: Annotation formats for CEGT-CCRL-TCEC

Post by Adam Hair » Thu Mar 13, 2014 10:22 pm

hgm wrote:Why does CCRL put the PV/alternative move within a comment? This seems a really bad idea. PGN has a standard for alternative moves/lines, and this deviates from the standard seemingly without reason.

Qb3 (Bg2) {-0.12/23 167s}

or

Qb3 {-0.12/23 167s} (Bg2)

would be much better, IMO.
We would have to ask Kirill. By the way, it is not an alternative move (as I understand the term), but rather the move predicted by the opposing engine. This is used for the ponder hit information we display at the web site.
I would be surprised if the EGT format was ChessBase. I thought the latter used this horrible [%emt...] format.
My only reason for guessing ChessBase is the German piece abbreviations in the pgn body.

User avatar
Kirill Kryukov
Posts: 492
Joined: Sun Mar 19, 2006 3:12 am

Re: Annotation formats for CEGT-CCRL-TCEC

Post by Kirill Kryukov » Fri Mar 14, 2014 4:54 am

Adam Hair wrote:
hgm wrote:Why does CCRL put the PV/alternative move within a comment? This seems a really bad idea. PGN has a standard for alternative moves/lines, and this deviates from the standard seemingly without reason.

Qb3 (Bg2) {-0.12/23 167s}

or

Qb3 {-0.12/23 167s} (Bg2)

would be much better, IMO.
We would have to ask Kirill. By the way, it is not an alternative move (as I understand the term), but rather the move predicted by the opposing engine. This is used for the ponder hit information we display at the web site.
I think the reason is that ChessBase GUIs used to put expected move in a comment. Chessbase GUIs were the most popular at the time CCRL was starting, so this convention was adopted for our unified output PGN format. This would not be hard to change, but it could cause some serious problems:

1. I believe that any PGN-processing software in use today knows about comments, however I doubt that every such software understands alternative move in parentheses. So moving alternative move outside commens may break some downstream analyses of our PGNs. This is made worse by:

2. Not all engines and GUIs produce sensible alternative move. Some use non-English piece abbreviations, some dump the whole PV, some print non-standard notation or just plain garbage. Putting such move outside comment would probably require us to validate that the alternative move is legal and in proper notation, which is not something I personally look forward to doing. (Although some very primitive normalization already occurs in our scripts, I can't guarantee that it's bullet-proof, and it does not even try to check the legality).

By the way, if someone was to produce a parser/normalizer/beautifier of PGNs, that would support the numerous non-standard notations and broken formats, we'd gladly consider incorporating it into our pipeline.

Best,
Kirill

User avatar
gbtami
Posts: 362
Joined: Wed Sep 26, 2012 11:29 am
Location: Hungary
Contact:

Re: Annotation formats for CEGT-CCRL-TCEC

Post by gbtami » Fri Mar 14, 2014 9:07 am

Adam Hair wrote:
hgm wrote:Why does CCRL put the PV/alternative move within a comment? This seems a really bad idea. PGN has a standard for alternative moves/lines, and this deviates from the standard seemingly without reason.

Qb3 (Bg2) {-0.12/23 167s}

or

Qb3 {-0.12/23 167s} (Bg2)

would be much better, IMO.
We would have to ask Kirill. By the way, it is not an alternative move (as I understand the term), but rather the move predicted by the opposing engine. This is used for the ponder hit information we display at the web site.
It _is_ RAV by pgn spec. http://www.saremba.de/chessgml/standard ... htm#c8.2.5

User avatar
gbtami
Posts: 362
Joined: Wed Sep 26, 2012 11:29 am
Location: Hungary
Contact:

Re: Annotation formats for CEGT-CCRL-TCEC

Post by gbtami » Fri Mar 14, 2014 9:12 am

Just notice, in 1. and 3. all black moves missing move number indication, so they are not valid export format .pgn regarding http://www.saremba.de/chessgml/standard ... htm#c8.2.2

Vinvin
Posts: 4244
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: Annotation formats for CEGT-CCRL-TCEC

Post by Vinvin » Fri Mar 14, 2014 9:19 am

Adam Hair wrote:... By the way, it is not an alternative move (as I understand the term), but rather the move predicted by the opposing engine. This is used for the ponder hit information we display at the web site.
I think it's the same thing : alternative move = variation = opponent's thinking =...

Post Reply