Very nice statistics, and most probably a lot of work behind. We did similar "studies" in the past. It should be nice if you add a column to that table, saying how many terms in the sequence the driver lasted. A good approximation is to multiply the breaking digits with a constant, assuming that the driver was there from the start. For example, D3 (2^3*3*5), with which I spent a lot of time in the past, grows like 2.8 terms per digit (iirc), while other (slower) drivers, like for example D2 (2^2*7) could grow as slow as 18 or 20 terms per digit (if no 3) or as fast as 2 to 4 terms per digit (if 3 and occasional 5).
Last fiddled with by LaurV on 20201209 at 02:45
