IMO there are many other issues with the current DB, including especially inconsistent/wrong/missing tournament names (in the Event field).

<IMO there are many other issues with the current DB, including especially inconsistent/wrong/missing tournament names (in the Event field).>

Many members of our site have been working on these issues- for years.

You can find some of these members here: Biographer Bistro.

============

Here is the webmaster's assistant who processes hundreds of database correction slips each month:

CG Librarian chessforum.

================

Here is the fruit of thousands of hours of volunteer labor to organize accurate game scores into accurate tournament collections, which have kibbutzing sections. This feature exists nowhere else on the internet except at Cg.com. It has never existed before in history, in fact. Now it does, and like all other features here, it can continually be improved as new historical information crops up:

Tournament Index.

===

Here is an ongoing project (still in its early days) to re-write the intros to the History of the World Chess Championship. So far, we have provisionally completed about 12 drafts from primary source material, all duly cited to the standard of academic history writing: WCC Editing Project chessforum. You or anyone else may inspect these drafts, which are marked "finished" in the profile. They still need to be polished and checked further before they can be used. We also need to finish researching and re-writing the rest of the intros.

###############

So the problems you speak of are being addressed.

I'm not sure what <Mr. Urcan> was suggesting Daniel *do* about all these problems- delete his website maybe?

Or maybe Daniel was supposed to personally examine every game that has been submitted to his site for a six month "waiting period" before accepting it?

I can assure you that other games databases are riddled with the same errors you find here. In fact, many, many of the games here come from those flawed databases.

The difference is this- to my knowledge, nobody is actually working to improve the historical accuracy of games in the "Chessbase Mega Database 2011."

If you really care about these database problems we are all aware of to begin with, perhaps you could bend your back to the grindstone and join us over here:

Biographer Bistro.

Yes I did you a disservice there, straying from your narrower point and addressing the contents of <Mr. Urcan's> article more than the contents of your post.

I suggest that you keep submitting correction slips. The Librarian really does examine and process them.

A few points about that:

1. The backlog can mean a slip is not processed for up to 3 months. So patience is necessary.

2. Be sure to include primary source verification for your corrections when you submit a slip.

3. In some cases, if your correction has not been processed for a much longer time, you can email chess@chessgames.com directly and this often helps bring attention to neglected slips.

<Karpova> Great work! You've probably noticed we've put a few of them online. I want to get our digital retouching down to a science before we go full-throttle with that project. On one hand we want to clean up the images, remove the yellowing and the artifacts from the opposite side of the pages. On the other hand, we don't want to compromise the data of the image in any way. The image on Ossip Bernstein for example has been greatly improved.

Wikipedia has it perfect and we would very much like to emulate their model from a technical standpoint. A good example is Richard Réti:

http://en.wikipedia.org/wiki/Richar...

You'll notice they include the accent mark. However, more interesting, if you do a search on "Richard Reti" it will take you to the correct page, with the accent, with a tiny note <Redirected from Richard Reti>.

The idea is, that at least 90% of users have no idea how to type a ñ or an é or an ä on their keyboard. They will just substitute n, e, and a. So to demand that the end-user types the diacritics is asking a bit too much. You'd be amazed at how many searches we get daily for "Bobby Fisher" [sic] … to expect them to type Réti instead of Reti would be absurd.

<You could use soundex for searches but still keep the names correct and consistent in the DB.>

We've considered Soundex, unfortunately it's not very good for our purposes. It was designed by Americans mostly for English names and doesn't work very well with slavic names. However you're definitely thinking in the right direction: we need to perfect our software so that end-users can type a variety of name variations, with or without diacritics, even without spelling the name quite right, all to search on the same person or persons.

<IMO there are many other issues with the current DB, including especially inconsistent/wrong/missing tournament names (in the Event field). I suppose many people care more about the moves and score than the names and event, but the latter are wrong much more often than the moves/score.>

Of course that's a product of the quality of the PGN which was originally received. The most glaring problem is that the Event and Site fields are often abysmal. A few pages back I examined some PGN from a brand-spanking-new tournament that listed Event as "?" and Site as "?". If people can't expect to get these details right for a 1-week old game imagine historians poring over these data 100 years from now!

Fortunately, as <jessicafischerqueen> has pointed out, the Biographer Bistro is taking great pains to normalize these data. The fruits of their labor is found at our Tournament Index. In cases like a Russian Championship or any event of such importance, these errors will fade away. It's just a matter of time before the championship you note will be incorporated into the Tournament Index at which point all of the event-tags will be consistent and the site tags will be corrected.

To an extent we've already done that, with the ability to change round numbers and years now in the hands of CG Editors. We're willing to extend such features, but care must be taken.

For example, many of the CG Librarian's correction slips are of the nature "This game cannot be 1-0 because Black is slightly better in the final position" or "Anand cannot have played this game because he would have been only 3 years old at the time."

Such slips are almost always correct, but they still are purely speculative. (In the first case: perhaps Black saw a ghost and resigned in error; perhaps Black's cellphone went off. In the second case: perhaps Anand did play the game, but the *year* is wrong instead of the player.)

My point is that publishing game scores is an inherently speculative enterprise. Since you can't get away from speculation either way, why not strive to be as accurate as possible?

I've mentioned before the idea of a logo you could use to brand games the scores of which have been changed due to speculation; something like "changed because our elves think this must have been played, but note that the original score was such-and-such..."

Take this game as an example-- I submitted a correction slip last May, and it hasn't been changed. Presumably the Librarian looked at it and concluded (correctly) that I was just speculating.

K Friedrichs vs B Schneider, 2001

The trouble is that "no" is too short a response. <cg.com> requires at least 5 characters in a response.

<Can I just ask what happened to the Slater - Tomlin game I uploaded.> It was a technical problem. Reviewing it now I was baffled at the error message "Invalid result." After staring at the screen for a few minutes I finally saw it:

[Result ""1/2-1/2"]

That extra " spoiled the PGN.

<Fixed and uploaded: James D Slater vs P C Tomlin, 1947>