31 May 2020

TWIC Data Through the Years

A couple of days ago, in TWIC and the Coronavirus, I wrote,

I wondered if I could somehow quantify the move to online chess. To answer that question, I developed the following chart based on recent TWIC data ('The Week in Chess' by Mark Crowther).

After a few paragraphs interpreting the data in the chart, I closed by saying,

Of course, this is all speculation that needs to be confirmed by reviewing individual TWIC files. Maybe I'll do that another time.

The 'TWIC and the Coronavirus' post was based on TWIC files starting with TWIC1300 (07-Oct-19). I went back to the earliest TWIC files and tried to apply the same techniques. The top-level structure of the files changed several times in the early years, but finally stabilized at TWIC0141 (21-Jul-97). The following charts show the TWIC numbers that had the largest and smallest file sizes for both the 'News' file and the 'Games' file.

The fields displayed in the chart are explained in the 'Coronavirus' post. One small caveat: TWIC supplied its own text version of the 'News' file through TWIC1004 (03-Feb-14). After that stopped, I continued creating a text version myself. It's possible that the sizes of the 'News' files before TWIC1004 are not comparable with later 'News' files. I couldn't see any significant difference in the before/after file sizes, but I didn't look very hard.

These numbers in the charts raise more questions than they answer. Will I find time for further investigations?

No comments: