Picking only the tail ends of the distribution also tells me you don't understand how the bulk of progress is made.

It isn't always Eureka moments but also a slow grinding away at assumptions to confirmations.