Anonymous partial solution
https://www.theverge.com/2018/10/24/...imathmystery
Quote:
there are no known applications for the formula, which isnâ€™t unusual in the field. It often takes decades, Pantone says, for formulas that are discovered in pure mathematics to make their way into realworld applications.

Well. A few seconds later:
Suppose one has software to test. In automated testing, how many case variations would constitute exercise for every possible sequence of discrete code sections between branches? Then along with computed bounds on running time of the code sections, one could estimate total test time. If any two or more sections have at least one variable in common, and modify that variable, one section can affect the behavior of the other(s).
Given that the bounds for 14 episodes were determined to be 93,884,313,611 to 93,924,230,411 possible sequential orderings, testability in feasible duration seems a strong argument for highly modular code.
Code does not have complete freedom for which section runs first, second, last, etc. and some sections may not be run at all in some cases. But the problem seems related.