chessgames.com chessforum (kibitz #22742)

So a game-with-position link could show as

Ponomariov vs L Bruzon, 2011 (move #37. ...)

I now discover that this is James McGrouther (aka James MacGrouther).

Not in the least.

Now if you typed "Alexander Alexandrovich Alekhine" it might create a new player record, until an admin merges the player records. And then the system gets smart and learns that "Alexander Alexandrovich Alekhine" is another name for player record #10240 and the next time it sees that it will know how to assign it.

Yes, if you send in PGN with this:

[White "Smith, John W."]

it will automatically get converted into this:

[White "John W Smith"]

<I'm sure it's been addressed before, but for players with the same name, is there any way of differentiating them upon submission?>

There is only one: the special tags <WhiteChessgamesID> and <BlackChessgamesID>.

For example, if you send us a game where White is Kurt Richter (and not Kurt Paul Otto Joseph Richter) then you can guarantee the correct association by adding this line to your PGN:

Damn it! I'm late.

<Penguin> A long time ago I speculated that there have been about 750K serious games of recorded chess and that's about how big a database should be.

The backdrop was that there was an "arms race" in which one database producer would try to top the other with sheer numbers. This led to gathering games from internet blitz servers, games between computers, amateur games, scholastic tournaments, and even worse. Heck, even duplicates were good for their purposes, because it meant "more games."

Chessgames did not want to go down that path.

http://norwaychess.com/

While on that topic, are you going to broadcast those games live? :)

There's also one more major tournament: the Capablanca Memorial.

http://torneocapablanca.inder.cu/

(busy time, I know)

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

It got me wondering how there's this:

<LIFETIME RECORD:

Classical games:
Edward G Sergeant tied Reginald Pryce Michell 1 to 1, with 1 draw.

*The figures above are based only on games present in our database which may be incomplete.>

and simultaneously this:

Let's call it <Date Rounding>, i.e. the tournament's <Round Numbers> can be determined simply by looking the the game's <Date>.

This assumes the tournament is a simple match, Swiss or RR, that doesn't need compound round numbers, and that all games are properly dated, with at most one round/day.

I'm belaboring the point to be explicit, but I think this is fairly familiar stuff.

Now, I have some, let's say, "candidate" tournaments for experimentation:

<
61997 2 12 ? Candidates Match: Aronian - Carlsen (2007)

61998 2 4 ? Candidates Match: Leko - Gurevich (2007)

61999 2 6 ? Candidates Match: Ponomariov - Rublevsky (2007)

62000 2 9 ? Candidates Match: Gelfand - Kasimdzhanov (2007)

62001 2 4 ? Candidates Match: Bacrot - Kamsky (2007)

62002 2 5 ? Candidates Match: Grischuk - Malakhov (2007)

62003 2 6 ? Candidates Match: Polgar - Bareev (2007)

62004 2 9 ? Candidates Match: Shirov - Adams (2007)

62061 2 6 ? Candidates Match: Aronian - Shirov (2007)

62061 2 6 ? Candidates Match: Aronian - Shirov (2007)

62062 2 5 ? Candidates Match: Bareev - Leko (2007)

62063 2 5 ? Candidates Match: Gelfand - Kamsky (2007)

>

Now, not every game is lacking a <Round Number>, but an automatic update could just save time and overwrite them all.

There is another tournament:

<62064 2 9 ? Candidates Match: Grischuk - Rublevsky (2007)>

which shows the way dangers can lurk...

Here there are multiple games on the same day (not allowed for a M-2), and there is also a "non-conforming" round number <R2 TB1> (or <RR2 TB1> using my typical notation).

I don't understand the head-to-head search.

I didn't specify anything beyond the two player's names. So the lifetime record should have been displayed.

Or maybe not?

If not, then why not?

Here's the search without that: