Another update on the items over the past couple of months that weren’t the subject of separate posts:
“Search games”: The search function for games in the IRLchess games archive has been rewritten. It’s now much faster–almost instantaneous, I think–instead of the previous wait of 30 seconds or so. See http://www.irlchess.com/cgi-bin/searchforgames.pl?query=brady or http://www.irlchess.com/cgi-bin/searchforgames.pl?query=daly for examples of how the new search works.
These two searches illustrate a point about the search as currently set up: if you’re looking for games by Stephen Brady, then searching for “Brady” matches games by any Brady: currently four games of Cormac Brady’s are included. But if you search for “Stephen Brady”, then (unfortunately) the search results will omit many of the games you’re looking for, as there are many games of “Brady,S” (no space) in the archive. That’s less than ideal, but the explanation is that TWIC gives names that way, and games extracted from older TWICs (before TWIC started adding FIDE IDs) show up like that. There are some improvements planned that address this issue, but in the meantime, I recommend trying a broader search.
Games archive: Games continue to be added to the archive. Apart from the ones from the World Youth Championship in Maribor and the World Senior Championship in Greece, the extra games that were not (yet) the subject of separate posts were:
I say “not yet” as there’s a post in preparation about one of these.
ECO classification: Here’s a topic that puzzles me slightly. I’ve added ECO classifications to all the games in the archives. Based on my own experience, this is one of the first items I look at when a new collection of games becomes available. (And you can search for any ECO code: e.g., typing B97 into the Search Games feature gives the Sicilian Najdorf Poisoned Pawn games.) The part that puzzles me is that ECO classification is not consistent across different sources. This would be understandable if the codes are typed in by players or tournament inputters, but even automated sources vary. For example, my favourite opening the French Winawer Poisoned Pawn 1.e4 e6 2.d4 d5 3.Nc3 Bb4 4.e5 c5 5.a3 Bxc3+ 6.bxc3 Ne7 7.Qg4 Qc7 used to be classified as C18 (by ECO, for example, and by Korchnoi’s monograph “C18-19” on the Winawer). Nowadays it’s often classified as C19: for example, this is the classification that TWIC uses. What’s the point of changing it? That simply makes finding games more difficult than it has to be, as you have to search for the openings you’re interested in under multiple codes. In any case, henceforth the policy here will be to convert all games to the default for the pgn-extract utility: I think this is the same classification that TWIC uses, and if TWIC isn’t the standard, what is?