The first rule of scientific fight club is that you have to agree on what you’re fighting about. A newly published debate on the merits of polarized training in endurance athletes, in Medicine & Science in Sports & Exercise, flunks this test. That’s actually a good thing, because the failure to disagree suggests that there might be some broad training principles that just about everyone in the field can get behind.
The concept of polarized training emerged about 20 years ago, thanks primarily to an American-born researcher in Norway named Stephen Seiler. It began as an observation about how elite endurance athletes in the modern era tend to spend their training hours: a huge amount of low intensity, a small amount of high intensity, and very little in the middle. That missing middle is why it’s called polarized: most of the training is at the low or high extremes of intensity.
Underlying this observation is the idea that you can divide training into three distinct zones. The easiest zone is anything up to your lactate threshold, during which you can probably talk in complete sentences. The hardest zone is anything above your critical speed, during which you can probably only gasp out a word or two at a time. The middle zone, between lactate threshold and critical speed, is often referred to as tempo or threshold training, and might allow you to speak in short phrases. (For more on how lactate threshold and critical speed are defined, see this explanation.)
Over time, the definition of polarized training has evolved and blurred. Matt Fitzgerald wrote a 2014 book based on Seiler’s research called 80/20 Running, in which the two higher zones are lumped together: the goal is to keep roughly 80 percent of your training easy and 20 percent of it hard. Other studies of elite athletes have turned up evidence of a slightly different distribution called pyramidal: easy training is still the foundation, but there’s slightly more of the middle zone than the highest zone. If a typical polarized distribution is 70 percent easy, 10 percent medium, 20 percent hard, the pyramidal equivalent would be 70 percent easy, 20 percent medium, and 10 percent hard.
This muddled terminology is the context in which Medicine & Science in Sports & Exercise’s debate takes place. Seiler teams up with a bunch of other big names in the endurance research field (Carl Foster, Arturo Casado, Jonathan Esteve-Lanao, and Thomas Haugen) to argue the proposition that polarized training is optimal for endurance athletes. Taking the opposite view are the equally credentialed team of Mark Burnley, Shawn Bearden, and Andrew Jones.
The Case for Polarization
The key plank in Team Polarized’s argument is the large number of observational studies of elite athletes in cross-country skiing, rowing, cycling, running, speed skating, and swimming that display either polarized or pyramidal training distributions. You’ll note that it says “polarized or pyramidal,” not just “polarized.” It turns out that the distinction between these two distributions is hazier than you might think.
For example, one study a few years ago that monitored the training of elite runners found they followed a near-perfect polarized distribution if you analyze the training zones based on running speed, but a pyramidal distribution if you base the zones on heart rate. An earlier study found that classifying training based on the overall goal of each workout led to a polarized distribution, whereas breaking it down by actual minutes spent in each heart rate zone produced a pyramidal distribution. So depending on exactly how you analyze the training, polarized and pyramidal may sometimes be describing exactly the same thing.
There have also been a half-dozen intervention studies in which athletes are randomly assigned to different training distributions for a week. For example, a 2007 study led by Esteve-Lanao compared five months of 80/12/8 versus 67/25/8 training for well-trained runners. The former group improved by 4.2 percent, the latter by 2.9 percent.
Foster and his co-authors spend some time wrestling with why polarized training might be superior to other approaches. In broad strokes, the goal of training is to accumulate as much adaptive stimulus as possible (i.e. get fitter) without triggering unwanted side effects like overtraining or injury. They argue that there are two main cellular pathways for boosting the mitochondria in your cells: one mediated by calcium signaling that responds primarily to high volumes of training, and the other mediated by an enzyme called AMPK that responds primarily to high intensity.
Polarized training, in this picture, is a way of accumulating lots of volume to max out your calcium-mediated gains with as little stress as possible, while including just enough intense training to max out the AMPK-mediated gains. The threshold zone, on the other hand, is stuck in the middle, not ideally suited to either pathway, and too stressful to allow you to rack up high volumes.
The Case Against Polarization
Burnley and his co-authors don’t think elite athletes’ training diaries can prove that any particular way of training is optimal. They’re right, of course. It’s easy to find examples of beliefs that were shared by champion athletes of one era—that drinking water during a marathon makes you slower, for example—and then rejected by the next generation. They’re also unconvinced that polarized training has any special ability to trigger calcium and AMPK signaling, an idea they dismiss as “rank speculation.”
But their biggest objection is that most of the observational studies of elite athletes actually show pyramidal rather than polarized distributions—at least “when training intensity is classified and quantified appropriately.” The same is true for some of the interventional studies, like the 2007 Esteve-Lanao study mentioned above, in which both groups are doing versions of pyramidal training. How could polarized training be optimal when all the supposed evidence is pyramidal?
This is where the debate goes off the rails. To the pro-polarization team, pyramidal is merely a variation on the general theme of polarized, as long as both adhere to the broader 80/20 principle of keeping most of the training in the easiest zone. When Seiler advocates polarized training, he’s talking about entire workouts: “I class a session as either hard or easy,” he told Runner’s World in 2019. “If I do an interval session, even though the effort and heart rate will fluctuate, it’s hard. If you run four times a week, no matter the length, if one run is hard then that’s a 75/25 split.”
To the anti-polarization team, on the other hand, it makes no sense to talk about polarization in the context of a two-zone 80/20 split. Polarization means avoiding the middle threshold zone—an impossible and nonsensical concept if there are only two zones.
I suspect everyone, including the authors of these viewpoints, would agree that arguments about terminology are less interesting than arguments about the concepts underlying the terminology. There’s a huge body of training data from elite endurance athletes that reveals some recurring patterns. Whether you analyze this data in a way that labels it polarized or pyramidal, the real question is whether this approach is truly optimal.
That question is particularly interesting at the moment, because there are some notable examples of current athletes who believe that threshold training—the forbidden zone, in a strict definition of polarized training—is actually the most important focus of their training.
Jakob Ingebritsen, who won the Olympic 1,500 race last summer at the age of 20, is the foremost proponent of what has come to be known as “the Norwegian model of lactate threshold training.” Marius Bakken, a former Norwegian Olympic runner, recently wrote a detailed account of how that model has evolved over the past two decades. Among the key planks: double threshold workouts (one in the morning and one in the afternoon) twice a week. Bakken even experimented with adding a midday session to get three threshold workouts in a single day, with the goal of accumulating as much time in that middle zone as possible. Olympic triathlon champion Kristian Blummenfelt reportedly uses a similar approach.
Even more recently, Swedish speedskater and double Olympic champion Nils van der Poel just published a manifesto outlining the training leading up to his 5,000- and 10,000-meter races in Beijing. It’s an amazing and idiosyncratic document for all sorts of reasons (he only trained five days a week… but occasionally completed challenges like a 100-mile run!). But what’s interesting is that he had a ten-week “threshold season” in which he racked up 1.5 to 2 hours of threshold training every day (not including his weekends off). He then transitioned to a “specific season” where he tried to do all his skating at race pace. Forget the polarized versus pyramidal debate—this guy is reading from a different book altogether… and setting world records in the process.
My own takeaways from this debate are somewhere in the middle. I don’t think there’s much evidence that threshold training is “bad” or should be avoided entirely. Whatever evidence exists is likely an artifact of the way the training is classified. I do think that the body of research on polarized training makes a strong case for the relative importance of accumulating lots of low-intensity training. In that sense, adding threshold training might be problematic if it comes at the expense of overall training volume—a trap that overenthusiastic recreational runners often fall into by pushing their easy runs harder than they intend to. But after watching Ingebritsen, Blummenfelt, and van der Poel demolish their by-the-book rivals, there’s no way I’d stick my neck out and declare any particular training approach as the one true path.
For more Sweat Science, join me on Twitter and Facebook, sign up for the email newsletter, and check out my book Endure: Mind, Body, and the Curiously Elastic Limits of Human Performance.