This (now greatly expanded) page has some additional points to note about the UltraCorrX-2018 database, including notes on some doubtful or corrected games.
Firstly we have aimed so far as possible for a standardisation of player names in accordance with ChessBase conventions.
For example, to facilitate searching, we use only the 26-character A-Z alphabet although some names with umlauts are converted as "oe" or "ue" (e.g.Moeller and Mueller instead of Moller and Muller) in German but usually not in Scandinavian names. It is advisable to use wildcards (*) when searching for those types of names.
Also we have tried to elininate hyphens from most names as they can cause problems in conversion to PGN, leading to part of the player's name sometimes being lost. We have also truncated many Russian patronymics as ChessBase does not accommodate very long names, and we have dropped titles such as Dr and Prof. in most cases.
ICCF databases put name prefixes such as De, Der, Von, Van at the end of the forenames, as library catalogues usually do. We prefer the ChessBase convention where these are part of the surname.
When we compiled UltraCorr-X last summer, our first update for many years, we were primarily concerned with adding new games from many sources which included the 2012 version of ChessBase's database. Unfortunately this led to many errors we had previously eliminated creeping back into the database.
We caught many mistakes before the release of the revised version in September, but now we found reviews we had written about the 2006 and 2008 versions of CB's database where numerous mistakes and spurious games were identified. In particular, games from a fictitious Munchhausen Memorial tournament (the name is rather a give-away!) and later "Pantos Memorial and "Maturin Gambit-ch" involving players whose names do not appear in connection with any other event. We believe these to have been deliberately inserted to catch plagiarists and we have now deleted them all. (ChessBase have always borrowed extensively from our databases so we don't mind returning the favour sometimes.) As we have not looked at their 2017 correspondence database we do not know how many of the old errors still remain there.
A major thing we did for the 2018 update was to go back to the BdF historic and server databases, frustrating and time consuming though it was. Maybe 10 per cent of about 60,000 games were new games or corrections for us (where we did not previously have the full orcorrect event or player names). That database includes over 10,000 games from ICCF European master class where the years were lacking and often the player names (usually forenames or initials) were wrong. By checking carefully against records of old ICCF events we have now been able to identify perhaps 95 per cent of these and made our best guess on the others.
There are also some interesting stories about players hidden in some of the annotated games. The originally Polish player Ernst Kocem changed his name to Kotzem when he emigrated to Germany and his games can be found under Kotzem.
Most curious, perhaps, was a case we solved many years ago. You will find under the name of Borloy Bata, Zoltan the games played by a man originally named Karoly Androvitzky. See his game against Baturinsky in the first European Team Championship in the 1970s. Puzzled by duplicate games in one of our early databases, we enquired of friends in Hungary. Ivan Bottlik informed us that this was indeed the same person; when he got divorced during the tournament he changed both his forename and his surname.
One of the most difficult types of error to eliminate is the over-the-board game that makes its way into correspondence databases (and occasionally the reverse occurs). One source of such errors is "CC" (for chess club) being misread by a database compile as "correspondence chess". In particular a lot of late 19thc entury games from Brooklyn Chess Club got into correspondence databases many years ago and some may still be around.
Another source of this error is when OTB games sometimes were cited in notes to postal games in Fernschach or other CC magazines and thus found their way into correspondence bases. Sometimes only the player concerned can inform us of such mistakes.
Apart from that, compared with the 2017 version, we have deleted the games Parr-Wheatcroft, Fairhurst-Alexander and Clemens-Eisenschmidt, all of which we had eliminated many years ago but crept back in last year via a ChessBase batch that somebody sent to me.
Also, the game Reshevsky-Leverett, 1982, is probably not a correspondence game. We have retained it, but with a note that it should probably be deleted after ctransferring to your OTB database. This Swiss tournament (CCA Summer International, New York) is mentioned in Gordon's book about Reshevsky pages 359-360 but this particular game is missing, so we cannot be certain.
All Reshevsky's CC games in that book are included here but I think Gordon did not know about Reshevsky's CC games with Mordechai Rechtman, at least some of which we also have.These were published many years ago in the now defunct online magazine Correspondence Chess News.
We also checked the book by Aidan Woodger for GM Reuben Fine's CC games so they are all included, we think, and most of the other McFarland collections were also consulted at one time or another.
We have retained Alekhine-Duhm, 1909 but are very doubtful about the date, occasion and identity of the players (possibly the champion's elder brother). The game is not in Skinner & Verhoeven'slarge book of the world champion's games and nor can we match it to any of A.A. Alekhine's postal tournaments (also researched by Charushin).
Please inform us of any more mistakes or doubtful instances of this kind that you may find.