Not according to this article, in which Joe Henderson states:
I told him that Whitlock’s runs all sound long to me. Before his 2:52 marathon of two years ago, all runs lasted two hours except the races, which could be shorter or longer. The racing itself served as “training” for speed or distance.
Sure, all runs are relatively long but that would mean all of his runs outside of races were surely under 20 miles.
Some more words of wisdom from that article and ones that give us a bit of a look into how long Whitlock’s runs actually were (sounds to me like we’re talking 15 mile range):
For two decades the long run has been king. We’ve been taught to make it good and long, approaching marathon distance or (in some programs) even beyond.
We’ve been told to do this weekly, or even two or three weeks apart, and to spend the time in between recovering with very easy runs and rest days. These long runs typically have at least doubled the runner’s everyday average.
There is another way — the about-the-same-every-day way. It requires two radical departures from conventional training wisdom:
1. Resurrect the old collapse-point theory, which guided marathon training programs of the 1970s and ’80s. This grimly named system held that you can run about three times your average daily distance before hitting a wall. Marathoners, then, need to up that average run to nearly nine miles to push the wall past the finish line.
2. Limit the length of long training runs to half-again the daily average. Adding 50 percent to an everyday eight- or nine-mile run would place the longest one at about a half-marathon.
Does a plan of shorter long runs and longer short ones work? It does for Ed Whitlock. He ran his fast marathons on daily training runs little more than half that length.