UTMB is one of the hardest events in all sports. 107 miles with 34,000 feet of climbing. It starts at 6 PM, and as the sun falls that first night, so do athlete dreams. The big, abstract goals get chewed up by the reality of high stakes on unforgiving trails.
Then UTMB proceeds to swallow souls.
But the real reason the race has trained away its gag reflex for soul-swallowing is that it’s competitive as hell. The list of professional athletes racing each year looks like a CVS receipt. Whoever writes the iRunFar previews is entitled to compensation for workplace carpal tunnel syndrome. When dozens of pro athletes on the start line are putting every drop of their souls on the line at the race, it turns into a physiological demolition derby.
The science on ultra training is still uncertain, with few comprehensive studies out there on the methods that work best. We’re in the process of figuring out what physiological variables correlate with top performance, like this 2022 study in the International Journal of Sports Physiology and Performance (summary article here), rather than settling on training approaches to optimize those variables. I bet we get there in 10-15 years, and I hope to help that growth from the coaching end, but anyone who says there is certainty is smoking some stuff I need to get my hands on.
That’s partially because trail running is really complex, mixing mechanical stresses with aerobic and metabolic demands in a way that means optimizing one variable can hurt another, especially over longer time horizons. And that’s partially because the sport is just recently getting competitive enough that it’s required to optimize every variable possible. Even at UTMB, the most competitive race in the world, the margins aren’t so narrow that athletes are desperately seeking every 0.1%. That’s why I love looking at other sports for lessons on endurance training at the peak of human capabilities. In cross-country skiing or cycling, physiologists are working with coaches and athletes to wring every drop out of endurance performance.
Who excels when the differences between the top performers are almost imperceptible to the naked eye? The answer to that question provides insights into the complex interplay between training theory, applied physiology literature, and actual real-world behavior patterns.
That brings me to one of the most fun studies of 2022. Published in June in the Scandinavian Journal of Medicine & Science in Sports, the study by Gabriele Gallo and collaborators examined 22 weeks of training for 3 male athletes that finished in the top 5 of the Giro d’Italia in recent years. This year’s race was a typical profile at 2,141 miles with 167,000 feet of climbing over 21 stages, pitting many of the best endurance athletes in the world against each other, with fame and fortune on the line. A race like the Giro requires moderately high max power, outstanding power at lactate threshold, stellar endurance, and even better fatigue resistance. To achieve the results, these 3 athletes had to tune their physiology to the absolute limit.
And we can learn really cool things at the limit.
It’s rare to get such a clear look at the training of the best athletes of the world in any sport, so I was downright giddy when I saw the study abstract. “This can give us lessons for trail and ultra running!” I thought, while printing off 6 copies (1 to read, 5 to gently caress when I get lonely at night). Given that these athletes are optimizing what the human body can do, this study promised a glimpse into what is often a black box: the complete training practices of the world’s best athletes.
I expected some major workout porn (hence the 5 extra copies), full of massive weeks, quad-busting workouts, and lifestyles that would make a monk grow hair just so they could pull it out. What I got was something much, much different.
Before getting to the lessons illuminated by those differences, some disclaimers. First, while cycling is an interesting analogue for any endurance sport, it has very little mechanical stress. My guess is the mechanical load from running down a few mountains at the start of UTMB is higher than the mechanical load from the entire 21 days of bike racing, depending on how you measure it. The cyclists are aerobic engines operating in a near-vacuum, whereas the runners are aerobic engines being tested in the middle of a hurricane.
Second, the bikers are optimizing performance for 4-6 hour days mostly, while the runners are looking at 20+ hours, followed by a couple weeks of hibernation. The physiology of fatigue resistance is likely similar at both durations and across multiple days, but it’s possible that different mechanisms will be rewarded.
Finally, cyclists wear shorts with perineum protection. This isn’t directly related to training, but I think we could learn a thing or two from how they watch their blind side.
6 lessons (7 if you count the perineum thing), let’s do this!
One: Athletes can excel off less aerobic volume than may be traditionally prescribed.
The three athletes’ weekly averages were far lower than I’d expect: 14.7 hours, 16.2 hours, and 19.7 hours. While that sounds high, cycling is non-impact, so there is no injury limitation. These are professionals who are being paid the big bucks, so if more training would be better for their aerobic systems, they’d probably do it. I had always heard 30+ hour training weeks had to be the norm at that level. This study slaps back at that notion.
The researchers had a mic-drop quote: “This is in line with the fact that even if training volume is considered a key factor of endurance training, increasing already high volume for long periods did not seem to further increase performance.” In other words, more is not better past a certain point. Given that running introduces more stress via mechanical demand, it’s likely that the optimal amount of time doing aerobic running training is much lower. That aligns with what we see in elite athlete practice, demonstrated in this 2022 study in Sports Medicine–Open, which found 450-700 hours of training per year common for world-leading runners, equating to 9 to 14 hours per week (summary article here).
For a trail runner averaging 12-minute miles on tough terrain in Chamonix, the 14.7 hour week would translate to 73.5 weekly miles, and at 10-minute miles, it would be 88 miles weekly. Even applying the cycling totals directly and not considering the impact stress, those numbers are lower than what some training approaches might expect from the world’s best ultra runners.
Two: Pulses of bigger training weeks may accrue adaptations without massive chronic stress accumulation across months.
While the overall training volume was not astonishingly high, there were some beastly weeks mixed in. The weekly max training was 27.7 hours, 29.3 hours, and 34.4 hours! It’s possible that the long-term volume totals are designed to best adapt to these big pulses, prioritizing the acute stresses of big workouts and weeks over the chronic stress of high monthly totals.
This aligns partially with what Jim Walmsley has been doing prior to this year’s UTMB. In a week at the end of July, he ran 35 hours with 164 miles and a whopping 64,000 feet of climbing. The week after was a step back to 17 hours, followed by 27 hours. We could be seeing that higher-volume pulses allow for adaptation without overstress for the best athletes in the world.
Three: Up to 90%+ of training can be easy for higher-volume endurance training.
Here’s the best part of the study! The researchers used a 3-zone model to quantify training intensity. Zone 1 (aerobic) was easy training, capped at 85% of 1-hour power. Zone 2 (threshold) training started there and went up to 1-hour power. Zone 3 was everything above that, including VO2 max intervals. The athletes spent a huge proportion of their time in zone 1: 83.6%, 86.7%, and 91.3% (for the higher volume athlete).
This aligns with studies on cross-country skiers and runners. In every endurance sport, it seems to be a prerequisite to spend 80%+ of time in Zone 1 in a 3-zone model, possibly to optimize metabolic processes and fatigue management via cellular level changes. And higher volumes require a bigger emphasis on easy volume.
You know how you hear that easy running is the gateway to fast endurance performances? That’s no longer up for debate–there is no substitute for going very easy.
Four: Pyramidal intensity distribution is how most pro athletes train in the real-world.
Intensity distribution refers to how much time is spent in each of those 3 zones. In Pyramidal training, athletes spend the most in Zone 1, followed by Zone 2, with the least in Zone 3. In Polarized training, athletes still spend the most time in Zone 1, but with a higher quantity of very hard work in Zone 3 and less moderate work in the “gray area” of Zone 2. All 3 athletes were strongly Pyramidal, with ~2-3 times more Zone 2 work than Zone 3 work.
That overlaps with long-term growth principles of aerobic development. The general rule is that Pyramidal training leads to slower-burn growth, but that can continue longer (possibly indefinitely). Meanwhile, the high-intensity emphasis of Polarized training may lead to rapid adaptation, followed by stagnation. That’s supported by a 2022 study on runners in the Scandinavian Journal of Medicine & Science in Sports, which suggested that stacking more intense Polarized training on top of aerobic-building Pyramidal training led to the best adaptations (summary article here).
So over 22 weeks, these athletes demonstrated Pyramidal training by necessity to avoid early stagnation. However, if you zoom in more closely, they did more Polarized training around races, showing that pulses of higher intensity may spur better aerobic adaptations from lower intensity work later. That likely shows that intensity distributions do not need to be set in stone in a block format–more intense pulses layered with less intense aerobic building can compound gains in both (aerobic development supporting better hard workouts and races, and those hard workouts and races supporting more economical aerobic development).
It can be helpful to have very hard workouts, training races, and training weeks, just make sure that consistent aerobic development maintains its foremost importance in growth across a training cycle (and over years).
Five: Rest days support long-term adaptation.
The athletes averaged between 0.77 and 1.31 rest days per week. It’s doubtful that these rest days were based on low motivation or limited time availability, but rather were about recovery and adaptation within bigger training cycles. Given that cycling has very low injury risk due to low impact forces, these rest days were likely geared toward the nervous and endocrine systems, allowing full resets that can support health and adaptation context.
If world-class cyclists take rest days, we can all take rest days.
Six: Strength work may be more optional than sometimes assumed.
None of the cyclists did any strength training. “That’s interesting,” I thought, especially since the research indicates that strength work helps cycling performance! And it has some overlap with what I have said about minimal-dose strength training for runners. Perhaps this indicates a new approach in endurance sports, focusing on maximum power in the activity and limiting extras to supportive work. I was so excited to tell you all!
But then I read the full paper. And the authors had something to say about that: “The reason why strength training was not performed was that the three cyclists were unwilling to perform strength training despite the coaches’ indications.”
OK, so maybe these superhumans are more like you and me than we might assume.
While the lack of strength training was not a decision coming from coaches, it does point out a fascinating point: what is demonstrated to work in studies may not apply to all athletes all the time. You’d think that if strength work was so beneficial, it would be a prerequisite to top performance in a sport with such narrow margins like cycling. The same goes for altitude training (one athlete did zero time at altitude) and tapering (all athletes did less extensive taper protocols than suggested in the literature). Are we seeing that something is missing in the literature?
I don’t think so. Instead, I think we’re seeing that training is always an N=1 experiment. The studies throw a bunch of N=1s together and can find significant associations, but there will be data points that fall away from the middle of the bell curve (unless it’s a universal intervention like the importance of easy training, which looks less like a bell curve and more like an orgy of dots within a small range).
By examining the N=1s individually, in running and across endurance sports, I think we can get clues that might unlock the next big advance in training theory. The jump from basic physiology to applied physiology is like jumping the Grand Canyon. The jump from applied physiology to training theory is like jumping the Atlantic Ocean.
And every one of these studies adds a few big pumps into our jumping shoes.
David Roche partners with runners of all abilities through his coaching service, Some Work, All Play. With Megan Roche, M.D., he hosts the Some Work, All Play podcast on running (and other things), and they answer training questions in a bonus podcast and newsletter on their Patreon page starting at $5 a month.