Michael Joyner, the Mayo Clinic physiologist and human performance expert who first floated the prospect of a sub-two-hour marathon back in the early 1990s, once distilled his voluminous training knowledge into a simple haiku:
Run a lot of miles
Some faster than your race pace
Rest once in a while.
A few years ago, he added some corollaries, one of which was “Make your hard days hard & your easy days easy!” Taken together, these guidelines describe a pattern known as polarized training. A large body of research pioneered by the Norway-based sports scientist Stephen Seiler has found that modern elite endurance athletes, whether they’re runners, rowers, or cross-country skiers, tend to organize their training along similar lines: huge volumes of relatively easy training; a smaller amount of very, very hard training; and almost nothing in the middle.
That pattern implies three different training zones (call them easy, medium, and hard), but descriptions of polarized training often lump the medium and hard zones together for simplicity. Matt Fitzgerald’s 2014 book about Seiler’s research, 80/20 Running, focuses on the idea of doing 80 percent of your running at an easy pace and 20 percent at a higher intensity. But that leaves an unanswered question about that middle zone, which corresponds to what are often called threshold or tempo workouts. Seiler’s research suggests you should avoid the middle zone, but there are some notable examples of athletes, like Paula Radcliffe, who view threshold running as a crucial key to success.
A new study in the International Journal of Sports Physiology and Performance, from a research group led by Phillip Bellinger of Griffith University in Australia, suggests a possible reason for confusion about the middle zone: it depends, in part, on how you measure it. Bellinger and his colleagues monitored the training of 14 serious middle-distance runners for eight weeks, using three different methods to classify the amount of time they spent in easy, medium, or hard training zones: running speed, heart rate, and perceived effort. The resulting training distributions were dramatically different.
The first question to consider is how you divide up the three training zones. There are a lot of different approaches, based on the lactate levels in your blood, your breathing rate, your ability to carry on a conversation, and so on. Most of them are pretty similar (though not necessarily identical). The approach used in this study was based on the relationships between the amount of oxygen inhaled, the amount of carbon dioxide exhaled, and the total amount of air being breathed. These relationships give you information about the balance between aerobic energy, which you rely on for easier efforts, and anaerobic energy, which is needed for harder efforts.
For practical purposes, you can think of two thresholds. The easiest training zone is below the first threshold, which corresponds loosely to the point at which your breathing begins to get noticeably heavier, making it difficult to speak in full sentences. The hardest training zone is above the second threshold, which is the point at which you’re out of breath and unable to speak more than a word or two at a time. The middle training zone is between those two thresholds. You can read more about these thresholds here.
The runners in the study completed an incremental treadmill test to determine their thresholds, and their heart rates and running speeds at each of these thresholds were noted in order to determine their training zones. For the eight-week training period, they wore a GPS watch and a heart rate monitor to track the time spent in each zone. They also rated each run or workout with a subjective effort rating between 0 and 10, with the easiest zone corresponding to 1 to 4, the middle zone 5 to 6, and the highest zone 7 to 10.
So, without further ado, here are the training distributions produced by each of those three methods:
The first thing that jumps out is that the ratings of perceived exertion (RPE) don’t show a polarized pattern at all. They find just 39.6 percent of training time spent in the easy zone, with 31.9 percent and 28.5 percent in the higher zones.
There are a couple of reasons for this. One is that a single effort rating for a given workout is too blunt an instrument. If I jog three miles warm-up, then run five times a mile as hard as I can with a quarter-mile jog recovery, then jog three miles to warm-down, I’ll rate the whole session as a very hard effort even though less than half the time was actually spent running hard. There are also some more subtle factors like duration: if I run 30 minutes at a moderate pace, I might assign that an effort rating of 4; if I run the same pace for two hours, I might rate it 6 or even 7, putting it in a different training zone even though the pace and heart rate were very similar. Perceived exertion is a really valuable tool that gives you important information about the stresses on your body—but it’s not the right tool to assess overall training distributions.
The distributions for heart rate and running speed are far more similar: they show 79.6 and 79.9 percent of time spent in the easiest training zone. If you’re just interested in a two-zone model (easy and not easy), both ways of measuring seem to produce equivalent results. (And the runners in this study, it turns out, did a pretty amazing job of nailing Fitzgerald’s 80/20 advice.)
But if you look at the two higher zones, there’s a difference. Based on running speed, the subjects spent 5.3 percent of their time in zone 2 and 14.7 percent of their time in zone 3. This is the classic polarized model, minimizing the time spent in the middle zone between the two thresholds. On the other hand, the heart rate data shows 17 percent in zone 2 and just 3.4 percent in zone 3. This is what’s known as a pyramidal distribution, with progressively smaller amounts of time in the higher zones.
The reason for this difference is pretty straightforward. When you accelerate to a fast pace, your heart rate takes a while to catch up to your pace. If you run 60-second interval repetitions in zone 3, it might take 30 or 40 or even 50 seconds before your heart rate gets into that highest zone for each interval. Interestingly, the same phenomenon was observed in a previous study of elite cyclists that compared training distributions based on heart rate, power, and perceived effort.
At the risk of making a stupidly obvious point, let me just note that the real-life training effects of a given workout are the same whether you analyze it with heart rate or running speed. So this data isn’t telling us that polarized is better than pyramid or vice-versa. It’s just telling us that your apparent training pattern may depend on how you measure it, which may be one of the reasons past research has found conflicting results about polarized versus pyramidal training.
In their conclusions, the authors of the new paper suggest that “integration of all three measures of training intensity” may be the way to go. That certainly makes sense if you have a full-time data scientist analyzing your training for you. In practice, if you’re interested in getting a big-picture look at your training distribution to see how it matches up to the ideals of polarized training, I’d suggest using running speed or cycling power or an equivalent. And when in doubt, if it all starts to seem a little too complicated or esoteric, you won’t go wrong sticking with Joyner’s advice.
My new book, Endure: Mind, Body, and the Curiously Elastic Limits of Human Performance, with a foreword by Malcolm Gladwell, is now available. For more, join me on Twitter and Facebook, and sign up for the Sweat Science email newsletter.