I mentioned probability theory.

Let’s consider Professor Hoffmann’s booklet *Bridge Whist* (along with its sister-publication, *Bridge*), and two specific periods of time — 1895 to 1899, and 1900 to 1903. (At the moment, I am not talking about post-1903 *Bridge* booklets, of which there are a great many.) The copies I am aware of can be listed basically as follows:

1895-1899: 2 copies

1900-1903: 12 copies

Or, if you change the spans by one year, they can be listed thusly:

1895-1898 (four years): 1 copy

1899-1903 (five years): 13 copies

Given the state of reality as we more or less know it, you don’t have to be an expert in probability theory to realize that it is likely that, in the whole world, there are far fewer copies of *Bridge* in existence for the years 1895 to (say) 1898, than there are for the period 1900 to 1903 (to take two equal periods, with a resultant 1-to-12 ratio of copies). You know intuitively that the distribution described above would be a very unlikely result if the copies of the book were distributed randomly in the two time-segments.

There are many reasons why this is a somewhat weak analysis. For instance, the sample size is quite small. Also, if the booklets that are “out there” truly exist in small numbers, then the removal of even one booklet can change the odds dramatically. Also, it is not clear what the exact starting and ending dates should be for the two periods. (For example, the 1895 copy probably did not appear on January 1, 1895.) And I am sure that there are other problems.

Nonetheless, here is one of the tentative conclusions that I draw from all this: There are about five times as many copies of *Bridge* (or *Bridge Whist*) in existence for the period 1900-1903 as there are for the period 1895-1899. We don’t necessarily know the numbers, but that seems a reasonable estimate of the *relative* number for the two time-periods. Maybe a more reasonable estimate would be a range of ratios: say, somewhere between 3-to-1, and 20-to-1.

Is it possible that there are really the same number of books in existence for each of the two time periods? Yes, of course — but it is extremely unlikely.