There’s a pace that feels productive. Not easy, not brutal. You get home tired, the file looks respectable, and you could do it again tomorrow. Most people do.
That pace is the most common way an endurance athlete stalls out for a year without noticing.
What the fast ones actually do
When Stephen Seiler and Glenn Kjerland followed junior cross-country skiers and counted where the training hours actually went, the distribution came out lopsided. About three quarters of the time sat below the first lactate turn point, at intensities the athletes themselves called easy. The hard work was there, but it was a small and very sharp minority of the total. Go looking in rowing, running, swimming or cycling and roughly the same shape keeps turning up.
Then people tested it head to head. Esteve-Lanao and colleagues took twelve subelite runners and split them for five months. One group put 81% of its time below threshold. The other doubled the time spent between thresholds. Both did the same small amount of genuinely hard work. Over 10 km of cross country, the easy-heavy group improved by about 35 seconds more.
Stöggl and Sperlich ran a bigger version, 48 well-trained athletes across four different distributions for nine weeks. The polarized group gained around 12% in VO2peak and 17% in time to exhaustion. The threshold group and the high-volume group barely moved.
The part people leave out
80/20 has become a slogan, and slogans get sloppy. There’s a live argument in the literature about whether the top of the pyramid should be a spike of very hard work or a band of moderate work. Burnley, Bearden and Jones published a flat rejection of the polarized model in 2022, arguing that a pyramidal shape, with more time just under threshold and less at the extreme, suits most athletes better. Foster’s group wrote back. It got spicy.
Read both sides and notice what nobody is arguing about. Every camp agrees the base is huge and easy. The fight is over the shape of the top fifth, not over whether the bottom four fifths should exist. There is no serious school of thought that says the answer is to spend your week at a steady, moderately uncomfortable pace.
How you count changes the answer
Here’s a wrinkle that explains a lot of the arguing. The same training week can be polarized or pyramidal depending purely on how you tally it.
Count by time in zone and a hard interval session mostly lands in the easy bucket, because the warm-up, the recoveries and the ride home are all easy minutes. Count by session goal, where the whole workout is labeled by its hardest intended work, and that same session becomes three hours of high intensity. Seiler’s group has pointed out that the two methods can move a distribution from roughly 80/20 to something closer to 60/40 without a single pedal stroke changing.
So when a study reports 80% easy, ask which 80% it means. And when you look at your own week, pick one method and stay with it, because the trend only means something if the measuring stick holds still.
Why it happens anyway
Because medium is emotionally efficient. Easy riding feels like wasting a morning, so the endurance ride drifts up ten watts, then twenty. Meanwhile the interval day starts on legs that are already two thirds spent, so the hard efforts land softer than the plan said. Both ends collapse toward the middle and you get a week of tempo wearing two different costumes.
The cost isn’t obvious for a while. You’ll hold decent form. You just stop getting faster, because nothing in the week is far enough from your normal to register as a signal.
Where medium is the right answer
It isn’t always wrong, and pretending otherwise is its own kind of sloppiness. Sustained work just under threshold is the specific stimulus for a set of real adaptations, and there are athletes and phases where it should dominate.
If your event is a long steady climb or a time trial, the demand is literally a sustained effort at threshold, and at some point the training has to look like the race. If you have six to eight hours a week rather than twenty, a strictly polarized week can leave you without enough total stimulus, and a pyramidal shape with more sub-threshold work tends to serve better. If you’re in the last block before a target event, specificity beats distribution theory.
The failure mode isn’t riding at medium. It’s riding at medium by accident, in every session, as the default result of never deciding. Chosen tempo is training. Drifted tempo is just erosion.
What to actually do
Pick a ceiling for your easy days and defend it. If you can’t get a full sentence out without a breath in the middle, you’re above it. Treat the prescribed pace as a ceiling rather than a target, which for most people means going slower than pride allows for about a month before it stops feeling wrong.
Then make the hard days earn the name. Two a week is plenty. If they’re genuinely hard they’ll feel like a different sport from the rest of the week, and that contrast is the whole mechanism.
OSO grades every session against your own critical power instead of a population chart, so “easy” is a real number for you rather than a feeling, and the weekly view shows where your time actually went. People are usually surprised the first time they see it. The gray zone is quiet, and it’s bigger than you’d guess.
Sources
- Seiler S, Kjerland GØ. Quantifying training intensity distribution in elite endurance athletes: is there evidence for an “optimal” distribution? Scand J Med Sci Sports. 2006;16(1):49-56.
- Esteve-Lanao J, Foster C, Seiler S, Lucia A. Impact of training intensity distribution on performance in endurance athletes. J Strength Cond Res. 2007;21(3):943-949.
- Stöggl T, Sperlich B. Polarized training has greater impact on key endurance variables than threshold, high intensity, or high volume training. Front Physiol. 2014;5:33.
- Seiler S. What is best practice for training intensity and duration distribution in endurance athletes? Int J Sports Physiol Perform. 2010;5(3):276-291.
- Burnley M, Bearden SE, Jones AM. Polarized training is not optimal for endurance athletes. Med Sci Sports Exerc. 2022;54(6):1032-1034, and the exchange that followed.