The site chess.com is located in Berkeley, California USA. That would put them in time zone UTC-8. They are 3hrs. behind chessgames(UTC-5) time.

Round 1 game: Rapport-Navara
1. Nf3 Nf6 2. g3 d5 3. Bg2 c6 4. O-O Bf5 5. c4 e6 6. cxd5 cxd5 7. Qb3 Qb6 8. Qxb6 axb6 9. Nc3 Nc6 10. d3 Bc5 11. Bf4 O-O 12. Nh4 Bg6 13. Nxg6 hxg6 14. Bd2 Rfc8 15. Rfc1 Ng4 16. Be1 f5 17. Rab1 Kf7 18. a3 Nd4 19. e3 Nb3 20. Rd1 Nf6 21. e4 Rd8 22. exd5 Nxd5 23. Nxd5 exd5 24. d4 Be7 25. Rd3 Na5 26. Bxa5 Rxa5 27. Rc1 Rd7 28. Rc8 b5 29. b4 Ra6 30. Rdc3 Bf6 31. R3c5 Rxa3 32. Bxd5+ Ke7 33. Rxb5 Bxd4 34. Rc2 b6 35. Rd2 Bc3 36. Re2+ Kf8 37. Kg2 Ra4 38. Bc6 Rd2 39. Rxd2 Bxd2 40. Rxf5+ gxf5 41. Bxa4 Bxb4 42. f4 g6 ½-½

Round 2: Navara-Rapport
1. e4 Nf6 2. e5 Nd5 3. d4 d6 4. Nf3 Bg4 5. Be2 e6 6. c4 Nb6 7. exd6 cxd6 8. Nc3 Be7 9. O-O O-O 10. h3 Bh5 11. b3 Nc6 12. Be3 d5 13. c5 Nc8 14. b4 a6 15. a3 Bxf3 16. Bxf3 Bf6 17. g4 N8e7 18. Bg2 g6 19. f4 b6 20. Na4 bxc5 21. Nxc5 Nc8 22. Rc1 N6e7 23. Qd3 Nd6 24. Kh1 Nc4 25. Bf2 Kh8 26. Rc2 Re8 27. Re2 Ra7 28. Rfe1 Rf8 29. Rf1 Re8 30. Ne4 Bg7 31. Bh4 dxe4 32. Qxc4 f5 33. Qxe6 Nd5 34. Qc6 Rc7 35. Qa4 Qd7 36. Qxd7 Rxd7 37. Bg3 Rc7 38. Kh2 Rec8 39. h4 Rc2 40. Rfe1 Bh6 41. gxf5 gxf5 42. Bh3 Nxf4 43. Bxf4 Bxf4+ 44. Kg2 Bd2 45. Rf1 Bc1 46. Rxc2 Rxc2+ 47. Rf2 Rxf2+ 48. Kxf2 f4 49. a4 Be3+ 50. Kf1 Bxd4 51. Bf5 e3 52. Ke2 Bc3 53. Bd3 Bxb4 54. Bxa6 ½-½

<cg.com: But when somebody continually harangues the same person, unprovoked, we make a move to stop it.>

Sure you do.

This is a little trivial, but I'm on the hound trail of hunting down <CG> updates, and the matter did come up.

OK, here goes, consider the <Anderssen // Staunton> games:

http://www.chessgames.com/perl/ches...

One game, <London (1851) R3.1 B40> shows up as 32 moves on the summary page. But consider the current version of the game:

Anderssen vs Staunton, 1851

This was on my radar as a recently modified game, where <CG> added the DSZ xtra move(s) <32...Qg4+ 33.Rg3 resigns.>

So, previously 32.Rf2 was the last move (as in Staunton's TB).

OK then, having 32 moves for the summary page makes sense with the old version of the game, but not the new (where it should be 33).

Don't believe this? Compare this game:

Staunton vs Anderssen, 1851

Moves = 30 as should be, with White making last play. So, the new <CG> game should have moves = 33.

Why doesn't it?

This has been long observed. Years ago the policy was to delete posts which were obviously confounded by a wrong score. Usually this involved a move that was ridiculous, a typo like Bxb3 instead of bxb3, and it was fairly easy to pick out the comments that were either confused or making light of it.

Later the notion came up of entirely deleting the whole game and starting the kibitzing from a blank slate. On a few very rare occasions when there is virtually no kibitzing and the score is a complete mess, we do that. But for 99%+ of the routine corrections that's just not an option.

Besides, we've learned that deleting posts can get people very irate, even when done with only the purest intentions.

So our new "Editor Notes" function was a kind of a patch to that scenario, so that if we have to change the score of a game we can at least be transparent about it, especially in situations where it's a very noteworthy change (i.e., most anything beyond an opening transposition). And if you go and look at that game, there are editor notes that even mention Staunton's tournament book.

<But notice should be provided in the comment stream, I strongly believe.>

In the form of a comment? Perhaps Sargon can actually leave a comment when something like that is changed. He only posts on very rare occasions but he'd happily be a chatterbox if instructed to.

It's hard for me to imagine that process of announcing corrections being fully automated though, and although you didn't specify that, I get the feeling that's what you are asking for. A CG admin's duties are so numerous and varied that most changes could not be explained in a succinct logfile type post.

For instance, often it's a matter of replacing faulty PGN with good PGN, thereby changing many moves at once, possibly even renumbering moves. We can't say succinctly "30...Bxb4+ replaced with 30...Rxb4" and leave it at that; it's rarely that simple of a situation.

Then one can backtrack to (mis)correct the score to match the commentary.

Plus, it leaves an audit trail in a place that seems natural, and easy to access.

The Editor Notes are currently unusable, imo. As you must be aware, without a filtering system, the S/N swamps any attempt to search for relevant updates.

Also- inserting correction notes into the comment stream can be done "a posteri" (assuming you logged changes to the SQL DB) - unlike the Editor Notes that only go back to recent times.

* * * * *

<It's hard for me to imagine that process of announcing corrections being fully automated though, and although you didn't specify that, I get the feeling that's what you are asking for. A CG admin's duties are so numerous and varied that most changes could not be explained in a succinct logfile type post.>

Yes, the auto-notice is what I'm asking for.

But, as you point out, it could be involved... still some notice is better than no-notice, that seems clear.

Most corrections I find (and I've found a whole lot by now) are usually one-movers, with the occasional Tpos.

I suppose a unix diff could handle the most general case, and even be semi-readable.

Leaving a link to an Ed-Note is another idea, but then, why not just insert the Ed-Note?

Let's reflect on a case where Ed-Notes might not need to go into the comment stream - i.e. Normalization changes.

Changing the Event tag is unlikely to affect any user comments. Changing the Site, on the other hand, might - although very rarely (esp. for <stonehenge>'s regularization changes).

Another case is changing the EventDate of a game. This should also have no consequence on a comment stream. Changing the Date might conceivably, although that would be rare again.

No solution is going to handle every case perfectly, but the current policy just leaves everybody in the lurch, it seems to me.

I also want to add another word in support of the EventSite tag - I'd really like the Site tags to be used for the specific game's site too.

I don't perceive it as an enormous problem, as we've taken at least two large measures to make the situation more manageable and transparent. The first is Editor Notes, the second is that handy little link that says "We have two scores for this game in the database, click here for alternate score."

I think the support should go to a GameSite tag.

<CG> can make the choice to not follow the standard of the <Site> tag - but it's a mistake.

It really isn't an issue for democratic vote at this point - it's a question of compliance with a pre-existing standard, one which <NIC>, <ChessBase>, and <365chess> chess all follow.

I've still yet to see any rationale presented why a GameSite tag doesn't satisfy the needs of users, or even a comment in the PGN.

Why is this?

* * * * *

BTW - if we were engaged in the initial design discussions for the PGN standard, I would almost certainly be 100% the EventSite tag, and appying the Site tag to the individual game location.

The trouble is, we weren't. And now we have practical considerations of backwards-compatibility to deal with. The GameSite tag seems to work well in this department.

And putting a comment in the PGN should satisfy any user viewing the game in a PGN viewer (needed for the Site tag anyways).

Look, Pope showed the right approach to this. When initially confronted with the correct meaning of the Site tag - he went back to check the PGN standard. He then acknowledged the proper usage of the tag. Of course, he stated that he would continue with his usage, even though nonconforming.

The difference is that he uses the tag for his own private DB. Anybody can do with their own DB whatever they want, of course.

The trouble is, <CG> is a *public* DB. As such, it really should embrace established standards, especially those that are widely used.

What did Steve Edwards have to say about that tag again, anyhow?

<8.1.1.2: The Site tag

The Site tag value should include city and region names along with a standard name for the country. The use of the IOC (International Olympic Committee) three letter names is suggested for those countries where such codes are available. If the site of the event is unknown, a single question mark should appear as the tag value. A comma may be used to separate a city from a region. No comma is needed to separate a city or region from the IOC country code. A later section of this document gives a list of three letter nation codes along with a few additions for "locations" not covered by the IOC.>

http://opensource.apple.com//source...

To me it is not clear at all that the Site tag was ever intended by Edwards intended to include multiple locations, in fact I would argue that it specifically is NOT for that purpose, because if it were, Edwards would have explained the proper syntax for separating them. (Slashes? Commas?)

The pragmatic issue here is that some software will not import a tournament properly if the Site tags are mismatched. And I can't blame them—imagine an event like "Grand Prix" that shows up with many different cities. The sane thing to do would be to separate them out into different events, not regard them as one gigantic long event taking place across the globe.

But on the other hand, some events like Bundesliga actually work very nicely as one giant page. At least I think they do, maybe others would disagree.

First - does Site apply to game or event?

The de-facto answer is that it applies to event. Look to 365chess and Chessbase and that's what they do(*).

Still don't believe me? Well, wiki happens to spell it out explicitly:

<2. Site: the location of the event.>

https://en.wikipedia.org/wiki/Porta...

What is wrong with putting the actual game's location in a comment?

Nobody, and I do mean nobody, has given me one single reason why this doesn't work for them.

<Seriously, why not?>

* * * * *

What's wrong with the Site tag being general enough to cover all locations...

e.g. <[Site "City_A / City_B CC-3"]>

or

e.g. <[Site "(Various) CC-3"]>

and then putting a

<{ @GameSite City A CC-3 }>

comment at the beginning of the game.

It keeps SCID/CB happy, and provides the user with the explicit location.

And, I might add, in a location that one could even see with pgn4web on <CG>, today!, unlike the Site tag.

Yes, we all agree the standard overlooked this obvious case, else we'd have the EventSite tag.

But we don't, and the horse is out of the barn. So, today we have the multi-site Site tag.

<in fact I would argue that it specifically is NOT for that purpose, because if it were, Edwards would have explained the proper syntax for separating them. (Slashes? Commas?)>

I don't agree with this logic. Absence of mention in the standard here should have mention elsewhere.

It was just overlooked.

Actually, what was overlooked was protecting name collisions and tournament grouping.

<The pragmatic issue here is that some software will not import a tournament properly if the Site tags are mismatched. And I can't blame them—imagine an event like "Grand Prix" that shows up with many different cities. The sane thing to do would be to separate them out into different events, not regard them as one gigantic long event taking place across the globe.>

OK, this is the key. In practice we have to find some strategy to group games into tournaments.

If you respect tournament grouping, then your hand is forced.

Tournament grouping is indeed driving the "de-facto" standard, and this applies to the Site tag.

I though we agreed on the importance of this when we began to realize the importance of normalization. It's the same issue.

<But on the other hand, some events like Bundesliga actually work very nicely as one giant page. At least I think they do, maybe others would disagree.>

The heuristic is to think of a tournament as being a tuple - i.e. (Event/Site/EventDate).

In practice this isn't even true, since the EventDate tag is often neglected (so much so that Tim Harding even bothered to write a long article about it).

<It could be that there is no steadfast rule here other than "Do as we see fit.">

There is a steadfast rule, which is used by Z-base, Chessbase, 365chess, and others.

All games in a tournament need to be normalized.

Hopefully <CG> will join this list.

Where is the compelling counter-argument not to??????