From the Gert Ingebreitsen interview 12/10/2021:
“A couple of years later I encountered a well-known Norwegian runner, I'm not mentioning any names, and he told me he's going to run a HM before running his next 5000 because that worked so well last year. Then I said," do you remember anything else you did the same period last year?" He replied, "Yes, I was sick for a week. I said "Well, that part will be hard to replicate?". After thinking for a couple of seconds, he said: "Yes, you're right about that". Then I said "It might not have been the HM that worked. Maybe it was the week of sickness that worked. How can you really know when it only happened once?" You have to do something more than once, to know empirically, because random events can affect the result and they are often hard to replicate. You have to make a plan that eliminates as much randomness as possible and makes the results predictable.”