ICU games archive update 2018

Seán Coffey


The ICU games archive has been comprehensively revised and updated. Summary of the main changes:

Extra games:

After the update the archive contains 28,660 games, compared to 21,142 in the February 26, 2018 version: an increase of about 35%.

Corrections:

About 300 games with various movetext errors were corrected (inconsistent or incorrect results, spurious moves from live boards or possible continuations appended at the end, garbled notes, invalid formatting, missing moves, and other game score errors). In about 200 of these, the ChessTempo software used for replaying games on the ICU web site generated a blank screen.

Duplicates, etc.:

Tags were added to identify 117 games that shouldn't be part of the archive (problems, duplicates, errors, and games with no Irish links), and in the near future these will no longer be included in the compiled archive, though the links to the games themselves will be retained.

Tournament names:

Tournament names have been expanded and made more consistent: "Irish Championship 2000" instead of "ch-IRL" and so on. In some cases a completely different tournament had been listed, and in others there was no useful tournament name ("Masters" or "2013" or even "?").

Player names:

Many (most but not all) player names have been expanded to provide first names. A few dozen cases of incorrect names were corrected (e.g., "Michael Kennedy" for Mel Ó Cinnéide).

Sundry additional game details (date, round, board, teams, event date, player titles, ratings, and so on) were added in mnay cases, and all annotations included in the November 19, 2017 version were retained (except that clock times and other live board artifacts were removed).

It would take too much space here to describe all the changes in detail. Instead a complete accounting of all differences between the update and the old version icu-20171119-163001.pgn (from November 19, 2017; missing only 24 games from this year's Bunratty Festival that were added on February 26) is contained in the file icu-update-with-comparisons-2018120-v0.pgn. This is a valid pgn file and should be playable using any pgn reader. It can also be useful to read it with text editors such as Notepad or Notepad++: the formatting is cleaner and it's possible to search for and count patterns across the entire file.

Here is one example that illustrates the format of the comparison file:

[Event "Oxford"] [Site "Oxford ENG"] [Date "1970.??.??"] [Round "?"] [White "Corden, Martyn"] [Black "Moles, John"] [Result "0-1"] [PlyCount "20"] [ECO "C15"] [Opening "French"] [Variation "Winawer (Nimzovich) variation"] [ICUid "15940"] [HashCode "1587bd42"]

{ Changes: "Site" changed: "?" -> "Oxford ENG" "ECO" changed: "C01" -> "C15" "PlyCount" (= "20") added "Opening" (= "French") added "Variation" (= "Winawer (Nimzovich) variation") added "HashCode" (= "1587bd42") added movetext change: was: (1. e4 e6 2. d4 d5 3. Nc3 Bb4 4. exd5 exd5 5. Qf3 Qe7+ 6. Nge2 Nc6 7. Be3 Nf6 8. O-O-O Bxc3 9. bxc3 Bg4 10. Qf4 g5 11. Qxg5 Bxe2 12. Bxe2 Qa3+ 13. Kb1 Ne4 0-1 ) --- end of changes }

  1. e4 e6 2. d4 d5 3. Nc3 Bb4 4. exd5 exd5 5. Qf3 Qe7+ 6. Ne2 Nc6 7. Be3 Nf6 O-O-O Bxc3 9. bxc3 Bg4 10. Qf4 g5 { 11. Qxg5 Bxe2 12. Bxe2 Qa3+ 13. Kb1 Ne4 } 0-1 The URL of this game is http://www.icu.ie/games/15940. This is captured in the pgn file via the new tag pair [ICUid "15940"]. For any game in which the movetext (the part starting with move 1 above) changed in any way other than line breaks and extra spaces, the old movetext is included in the comment preamble. For the game above, the change is that the game actually ended after 10... g5, and the extra moves should have been shown as a possible continuation.

(The update contains many pgn tags (ICUid, ICUType, and others) that will not be retained if the pgn file is converted to .cbv format. If you convert to .cbv format and then convert back to pgn, you will find that about half the tags are deleted, and sometimes that new tags are added. However the games themselves will be playable.)

The changes take up over 250,000 lines, but the vast majority are minor: repetitive (e.g., event names, player names) or automatically generated (e.g., ECO code, opening, plycount, player name format).

Acknowledgements:

The update also incorporates an earlier significant overhaul prepared by Tim Harding that added 1,288 new games, in addition to many changes of the types listed above. David McAlister helped resolve several particularly awkward discrepancies. And Jonathan O'Connor made several very useful changes to the database storing the games archive on the ICU web site.

Finally, it's virtually certain that there are many remaining errors lurking in the update. If you notice any, we'd be interested in hearing from you: please send corrections, and any new games, to [email protected].


Created 2018-03-16 ◦ Last updated 2018-03-16 ◦ Editor JMM


New Search
© 2004-2024 Irish Chess Union ● Contact UsPrivacy Policy