The guy who made the soon to be released film on the sub-2 attempt said this to Runners World last week, "I watched him (Kipchoge) do a 40K as his training run, and if he had done the extra 2K he would have clocked in at a 2:08 marathon, and this is at altitude as part of his training."
While reading it, I'll admit my first thought was, "Is that even possible to run that fast at altitude?"
I like to let the stats speak for themselves.
So let's do the math.
2:08 could be 2:08:59 which is 4:55.2 pace. World record pace is 4:41.4. So he ran 13.8 seconds slower than marathon WR pace for slightly less the distance.
The NCAA conversion on that performance would be mind-boggling. However, many Kenyans (but not Paul Kipsiele Koech) are very good at running at altitude. Remember, Geoffrey Kamworor has run 27:11 for 10,000 at altitude.
27:11 is 4:22.5 pace. The WR for 10k is 26:17.53 which is 4:13.8 pace. So there is a difference of just 8.7 seconds per mile.
So it does seem doable that someone could run 13.8 seconds slower than marathon pace for 40k in practice.
So the verdict of first ever LetsRun.com fact check is : Seems plausible.
Quote came from this article: