[Continued]
That there is a hard core of solvers is also suggested by an analysis of the numbers of solvers who have submitted more than 25 puzzles each year. Since records of this began in 1992, that number has remained roughly steady at an average of about 430, although the number may have increased slightly in recent years (average of 490 in the last 5 years), and the regression line shows a small gradual increase over the whole period. During the same period, the number of entrants who have submitted only one solution during the year, although highly variable, has fallen on average.
I suspect that all this means that overall the puzzle is slightly more difficult to solve today than it used to be, and has therefore become of interest to fewer people than before. However, because the available aids, if you have them and are interested, mitigate this increase in difficulty, the coterie of regular solvers has increased a little.
This is consistent with my impression that setters have responded to the availability of solving aids by putting higher barriers in the path of the solver. For example, it has become more common for them to use definitions that differ from the exact (i.e. searchable) wording in Chambers, and there has been an increasing use over the years of misprinted definitions and other obfuscating techniques.
Those who would like copies of the graphs that I have prepared based on the complete data set should e-mail me at
[email protected].