Two weeks ago, in Korchnoi's Career 1945-1977 (on Paper), I extracted an index of Korchnoi's games for the period 1945-1977 from the book 'Korchnoi's Chess Games' by Levy & O'Connell. Last week, in Korchnoi's TMER 1976-2000, I compared games for the period 1976-2000 from two PGN collections compiled around the turn of the century -- one by UPITT and one by GMchess.
Of course, the two PGN collections also include games going back to 1945. How do these three sources, a book and two files, compare in their coverage of Korchnoi's pre-1978 period?
The chart on the left shows a count of games played in a given year for each of the three sources, plus an average of the three sources. For any particular year, the counts are roughly equivalent.
How does the data compare for a specific year? The following table shows events for the year 1977 from the three sources.
|GMC||1977||Ch Germany (team) (South-west) 1977/78 * Germany||1|
|GMC||1977||Ch Netherlands * Leeuwarden (Netherlands)||13|
|GMC||1977||Ch World (match) (cand.) (1/2) * Evian (France)||12|
|GMC||1977||Ch World (match) (cand.) (1/4) * Ciocco (Italy)||13|
|UPI||1977||Belgrade cm f||17|
|UPI||1977||Ciocco cqf ,CAND||4|
|UPI||1977||Evian cm sf||12|
|UPI||1977||Il Ciocco cm qf||8|
|UPI||1977||Leeuwarden NED ch||11|
While the data again looks comparable, it's only because I eliminated 75% of the 'events' for UPITT, where the PGN headers haven't been standardized. Even for the six remaining events, Ciocco and Il Ciocco identify the same Candidates match.
To make progress for the period 1945-1977, I need to work only with the Levy & O'Connell material (on paper) and the GMchess data. The UPITT data can be ignored for now.