> isn't it just as plausible that this median would increase? Every sixth number will have a factors 2 and 3. (so 3 is the second prime for 16.666% of numbers

> isn't it just as plausible that this median would increase? Every sixth number will have a factors 2 and 3. (so 3 is the second prime for 16.666% of numbers) Every 10th number will have factors 2 and 5. Every 15th number will have factors 3 and 5. Of course then you have to subtract out every 30th number which contains a 2, 3 and 5. They ignored duplicates and it would have been nice if they had defined "first" and "second" properly. Also note that a prime counts the 2nd factor as "infinite", adding one to the "above 37" bucket. Anyhow, you add those up and it's easy to at least see that in any range of numbers it very quickly approaches 50%. Keep going to write out the Venn diagram and it happens to come out at 37. It's low enough, I assume it could even be done by hand.