In Alex Hutchinson's latest report on Nike's attempt to have a man run a marathon in under two hours--which you should read, like everything Hutchinson writes, for its clear explanation of exercise science--is the detail that Eliud Kipchoge, the greatest marathoner of all time, has run on a treadmill two times ever.

Lelisa Desisa and Zersenay Tadese, the other two men attempting to go sub-2:00, delivered tons of useful information in the lab. For Desisa, the measurements of his glycogen stores led to the conclusion that he wasn't fueling enough during races. For Tadese, his unexpectedly low lactate threshold meant that "has all the speed he needs but should work on sustainability," doing tempo runs and longer intervals. (Of course, sometimes science says the obvious. Desisa dropped out of the New York City Marathon three months ago, something that could be obviously explained by inadequate fueling. And Tadese has the half marathon world record but has never broken 2:10 in the full, so yeah, he needs to work on sustainability over speed.)

But Kipchoge's lab testing was of...less use. University of Exeter scientist Andy Jones basically refuses to predict Kipchoge's chances of breaking 2:00, because, as Hutchinson writes:

"Kipchoge, he said, was difficult to predict, because his visit to the lab for testing had been his first time on a treadmill. When I visited Nike HQ in Oregon last year, I saw Kipchoge's second treadmill testing session, and I can attest that he looked extremely uncomfortable--like Bambi on ice. So his lab numbers probably don't reflect his actual potential."

Treadmills are bad.