Very interesting and I listened to most of it, but kept wondering about the times of year when people are more or less likely to make babies. Sure, mathematically, there can be the small number, or ten times that number, but another reason for the latter number is that the author - and others to which it applies - has an unpopular birthday (arguably). The most popular birthdays are in mid-September. So I'd think one should try to weight mid-september in the calculation, against the least popular, less weighted period, which tends to be in late December or early January, and all other periods would have the same relative likelihood for a shared birthdate. This would add two variables to the calculation, each with more and less weight for thier likelihood. And with those three variables, one might have a more accurate estimate between the lowest calculation and the highest calculation - or arguably - half the value of the lowest calculation and the highest calculation, but I've not tried that calculation. My argument is that the two extra variables (as divisions of 36, if we divide months into 3) could help with the accuracy of the calculation.