The Science On What Variables Predict Trail Running Performance

Published April 19, 2022 07:49AM

A wonderfully fun study published in February 2022 asked a provocative question: what are the physiological determinants of trail running performance? The authors had me hooked in the abstract. So I clicked “Access Full Text.” I was blocked as if the journal was Dikembe Mutombo.

No worries, I thought. My co-coach/wife Megan has that sweet university access and I imagine that spousal confidentiality allows us to share sexy secrets and/or articles. BUT! Megan was denied as well, this journal filling up a moat with alligators that hate the democratic dissemination of science.

So … I did it. I clicked “Buy” and paid $24.95 to Human Kinetics to purchase a single article. And yes, that’s the same company that published the book that Megan and I wrote, which made them lots of cash. I am a sucker for science, who also sucks at business.

But it was the best $24.95 I have ever spent, because the article is amazing and adds texture to a field of study that has exploded over the last few years. The general outline of the studies is simple: they gather athletes that are doing a certain race, have those athletes do a series of baseline tests for physiological attributes and fitness, and then look for correlations between those physiological metrics and subsequent race performance.

It’s the sexiest science because of that race performance variable.

Often, studies use protocols to determine the fitness metric that encompasses the dependent variable. An example might include a time to exhaustion test, where athletes run on a treadmill that gets progressively faster until they fly off the back and through the lab’s drywall. But those protocols have limitations, particularly for trail running, where performance is messy and complicated. These studies use the lab tests for inputs, and race performance for outputs, bringing theory and reality together in a hot tub of scientific bonding. Megan and I talk in more detail on our podcast here.

Let’s break down some of the results, starting with the study that costs more than six boxes of Chocolate Chex. The February 2022 study published in the International Journal of Sports Physiology and Performance recruited 75 runners participating in one of the UTMB races (OCC at 55 km, CCC at 101 km, and UTMB at 145 km). Of the 75 athletes recruited, 54 finished and were included in the analysis (24 in OCC, 16 in CCC, 14 in UTMB). In the weeks before the race, the researchers measured:

It’s so much science that it risks losing its appeal, like if all you see is butts everywhere, butts end up being kinda normal. Actually, scratch that. There is no such thing as too many butts or too much science.

Race results were standardized based on performance relative to the winner, allowing researchers to find what variables correlate with higher relative performances.

At the 55 km OCC race, performance correlated with VO2 max (r = 0.86), peak velocity on the treadmill test (r = 0.85), and fat/carbohydrate oxidation at 10 km/hr (r = 0.59). Knee extensor strength correlated as well (r = 0.58). Speed, strength, and aerobic capacity ruled the day.

At the 101 km CCC race, performance correlated with VO2 max (r = 0.92) and peak velocity (r = 0.90), but not with oxidation rates. Like in OCC, knee extensor strength correlated too (r = 0.55), with an added correlation with the bike sprints (r = 0.69). Speed, strength, and aerobic capacity continue to rule the day.

At the 145 km UTMB race, performance correlated only with peak velocity (r = 0.62). Chaos reigned supreme.

Interestingly, “there was no relationship between performance in any distance and body mass index, body mass, or total fat free mass.” Focus on what the body does, not how the body looks. Finding your strong is key.

What a cool study!

To summarize the main findings: peak velocity achieved on the treadmill test was the only variable to correlate with performance at all distances, with VO2 max seeming to be important as well, and strength-related variables having some mid-level correlation at times. One of the more interesting findings to me was that metabolic variables seemed to matter most at shorter distances. My theory is that those shorter distances are more intense, thus they are closer to aerobic threshold, so metabolic efficiency plays a larger role. Meanwhile, in very long events, all athletes are way below that metric or they DNF.

This study has some limitations, particularly the removal of DNFs from the data analysis. In addition, correlation is not causation, so we could be seeing proxy variables for the real drivers. My biggest concern would be that a number of other variables might show similar correlations if we measured them. Some I’d be curious about: training approach, athletic history, treadmill hiking speed, muscle damage from downhill running bouts, fatigue resistance, gastric emptying rate, and about a million others.

But the researchers did such great work, and the findings about speed and strength are especially fascinating when viewed together with some other studies in the field. For example, a 2019 study in the Journal of Human Kinetics looked at 25 runners before a 31 km or 21 km trail race. Those researchers found that lactate threshold speed (at 4 mmol lactate, which you can ballpark at around 1-hour effort) correlated most strongly with performance. While the 2022 study didn’t measure that variable directly, I’d bet $24.95 that it would show a similar relationship as peak velocity.

A 2018 study in the International Journal of Sports Physiology and Performance evaluated 9 competitive male trail runners prior to a 31 km race, finding that speed at lactate threshold and running economy at 12 km/hr had the highest correlation with performance. A 2020 study in the International Journal of Sports Medicine broke it down between men and women prior to a 107 km ultra. The general findings were that peak oxygen uptake (more important for women) and peak speed (more important for men) had the strongest correlations with race performance, with additional significant correlations with speed at aerobic and lactate thresholds, plus metabolic variables.

Whew, that’s all the studies. You can put down your reading glasses, and take a satisfied sip of your chamomile tea.

JUST PLAYIN! Chug some Four Loko and let’s dive back into some studies that help bring it all together.

Even More Studies

A 2021 study in the International Journal of Sports Physiology and Performance found that peak velocity had the strongest correlation for athletes doing 50 km and 50 mile races, but there were no correlations for 100 miles. Those authors had a sexy conclusion: “While classic determinants of running performance, including cardiovascular health and running fitness, predict 50-km trail-running success, performance in longer-distance races appears to be less influenced by such physiological parameters.” That reminds you a bit of the 2022 study, right?

A 2021 review will put us out of our Four Loko-induced fugue state as it summarizes the state of knowledge on physiological indicators for real-world trail running performance. While the review acknowledges that trail running is less predictable than road running, it found that the following associations (Note: every variable wasn’t measured in every study, and standard deviations were all over the place, so this leaves out some important nuance and shouldn’t be viewed as a ranking of importance):

I love these studies so much that any of the articles could be a third in our marriage if they showed up to our door with a vaccination card and free journal access.

There are three takeaways for trail runners.

One: Peak velocity, VO2 max, and lactate threshold matter, even in long mountain races.

Across most studies, there seems to be a correlation between upper end output and trail race performance. That overlaps with what we see in coaching–an athlete that develops their speed will not nail every race, but they give themselves a much higher chance of long-term growth that makes their cone of probability encompass more great results.

To throw a turd into the punch bowl, here is where the study protocols get tricky. In the 2022 study, peak velocity was measured at 12% grade. In other studies, it’s on level ground. In some, the authors measure velocity at VO2 max; in others, the athletes can keep going until they make a runner-sized hole in the drywall. Perhaps most disconcerting, maybe these variables are just measuring adherence to a training schedule, making us think we’re measuring speed when we’re actually measuring diligence with training more generally.

Let me fish that turd out of the punch bowl and throw it in the trash. Running economy on level ground correlates with running economy on uphills, so working on speed on both flats and ups should work in every direction. In addition, top-end running economy is correlated with submaximal running economy in athletes over time, and there is a strong argument that continually developing speed is the best way to improve over multiple years.

My big guidance for athletes: get fast with strides (1-3 times per week) and short intervals on flats and ups (usually once per week), stay fast year round (with some down weeks mixed in), and layer that on top of continual base building with long runs that are specific to the terrain you race on.

Two: The longer the ultra, the greater the uncertainty. Control the controllables.

The breakdown of correlations at 100 miles reminds me of the quote from schemer Littlefinger in Game of Thrones, said in response to a moment of upheaval. “Chaos is a ladder.” While chaos creates unpredictability, it’s also an opportunity.

Speed still matters, so keep the principles above in mind. But also emphasize ways to improve fatigue resistance, particularly with downhill running resilience, uphill hiking, and fueling. You have to be a bit more willing to accept that results won’t flow as directly from training, but that’s also part of the fun. Climb that chaos ladder.

Three: Focus on strength, health, and embracing your full athleticism

While I focused the summary on running training metrics, some studies found cool correlations with things like blood pressure and knee extensor strength. When the events get wacky, an athlete is not calling on their performance in a 15 x 1 minute fast/1 minute easy interval workout. That type of workout may improve VO2 max, lactate threshold speed, and peak velocity, but it will have almost no direct impact on trail race performance, as much as my speed-and-strides-loving heart would love to say otherwise.

Instead, speed development through well-rounded training feeds into trail and ultra performance indirectly. Those cells-and-systems level stimuli from consistency and workouts are processed by the nervous and endocrine systems, layering on top of life stress to impact the aerobic and musculoskeletal systems via longer-term adaptations. If any link in the physiological chain is weak, an academic model of how fitness impacts trail performance may break down.

So get strong and stay strong with simple routines. Manage stress and don’t train yourself into a ditch. And most of all, make it fun. I love these studies with all of my heart, but what I’d love even more is to add another variable. At 4, 8, and 12 weeks pre-race, let’s ask the athletes: “Are you enjoying the process?” And as a follow up: “How cool is this shit?”

Let’s put their responses on a scale of 1 to 10. Let’s correlate it with race results.

I bet there’d be some association, if not in a single race, then over a few years. And even if the P-value is too high to make it into the abstract and get $24.95 from a poor sucker, that’s OK too.

Because the athletes that love the process are doing it right… whatever “it” we are measuring.