Physiological resilience: A key component of marathon and ultramarathon performance

Why do some runners slow down dramatically in the final miles of the marathon? Is it just “the wall”—running out of glycogen—or is there a better, more comprehensive way to understand the late-race demands of the marathon and ultra-distance events? 

This article takes a look at a newly-proposed component of fitness that elegantly explains the late-race challenges you face in the marathon and ultramarathon: resilience.

Physiological models of running performance have historically focused three distinct factors, or “components of fitness”: VO2max, maximal metabolic steady-state (SSmax, often approximated using lactate threshold), and running economy. 

Researchers trying to model performance in the 800m or 1500m will sometimes incorporate anaerobic energy reserves as well, but for long-distance events, the “big three” of VO2max, SSmax, and running economy have been the central focus of physiology research since the 1970s. 

However, resilience is a newly-proposed “fourth dimension” for endurance performance that represents how well you can resist deterioration in the other three components of fitness over the course of a long race like a marathon or ultramarathon.

(Resilience is sometimes called “physiological resilience” to differentiate it from the more general psychological sense of "resilience" meaning toughness and grit. In the cycling and ultra running world, resilience is sometimes called durability, though that term is sometimes used to describe how well someone can handle high training volumes from day to day and week to week.)

Resilience is still very much an emerging area of research, but I’ve found it to be an extremely useful concept when planning training for marathon and ultra-distance events, and for problem-solving after a disappointing race.

Before we dive into the science of resilience, it’s worth doing a quick overview of the classic “big three” components of fitness: VO2max, SSmax, and running economy. 

VO2max: maximal oxygen intake

VO2max represents the maximum volume (V) of oxygen (O2) that you can consume per minute. 

VO2max is assessed with an incremental treadmill test ideally lasting 12-16 minutes or so, and is measured either as “absolute” VO2max, in liters of oxygen per minute, or more commonly as relative VO2max, which is absolute VO2max divided by body mass. 

It’s useful to think of any given rate of oxygen consumption (i.e. a given VO2) as a rate of aerobic energy production. Your VO2max, then, represents your highest possible aerobic energy production rate, or equivalently, your maximum aerobic power output ("power" in the sense of metabolic power). 

Steady-SSmax: maximal metabolic steady-state

Your maximal metabolic steady-state, or SSmax (“steady-state max”), represents the highest aerobic power output that you can sustain at a metabolic steady-state. 

What’s a metabolic steady-state look like? It means that, as you run at a constant speed, you do not see significant changes over time in your VO2, blood lactate, intramuscular acidity levels, or intramuscular phosphocreatine concentrations. 

In contrast, if you are not at a steady-state, all of these biomarkers diverge: VO2 rises inexorably towards your VO2max, blood lactate spirals out of control, intramuscular acidity skyrockets, and phosphocreatine concentrations plummet.  

In casual conversation, we often talk about SSmax as a speed (e.g. “repeats at lactate threshold pace”), but for modeling performance, it’s more correct to think of it as a rate of aerobic energy production. Indeed, SSmax and its surrogates are typically expressed in research as a percentage of VO2max.

Measuring SSmax with lactate testing or critical speed

You probably have not heard the term “SSmax” before. However, you’ve almost surely heard of lactate threshold, critical speed, or maximal lactate steady-state (MLSS), which are three different tests that can be used to estimate SSmax. 

The two competing gold standards for SSmax are maximal lactate steady-state (MLSS) and critical speed (CS, a.k.a. "CV" or "critical velocity"). Lactate threshold—more properly called LT2 for “second lactate threshold”—is a less accurate estimate of SSmax that, unlike MLSS and CS, can be measured in a single lab visit. The simplicity of LT2 makes it by far the most popular way to estimate SSmax.

You can express maximum metabolic steady-state as a VO2, but it’s more useful to express it as a percentage of your VO2max. A typical value might be 80-85% of VO2max for a reasonably well-trained runner, though elite marathoners are north of 90%.  

Running economy

Most people have a vague and mostly-correct impression that running economy represents a sort of “efficiency” of your running. Indeed, a good way to think of running economy is as a conversion rate between metabolic power output and forward velocity. 

If we want to be more precise, we can express running economy as the amount of oxygen you consume over the course of one kilometer, normalized to body mass.

Running economy is often expressed per kilometer, not per minute or per second, because it makes the measurement independent of speed. 

Running economy is almost constant across different speeds; the plot below shows running economy data from a 2018 study in both elite runners (sub-31:00/sub-35:00 10k for males/females) and recreational runners.

The elite runners, not surprisingly, tend to have better (i.e. lower) running economy. There is a bit of a U-shaped curve, but the metabolic cost of running is mostly flat, especially for faster speeds. Note that this plot measures energetic expenditure, not oxygen cost, which is almost the same thing. 

Combining the big three

If we’re careful with our units, it’s possible to combine the big three physiology metrics to get a prediction of your fastest aerobically-sustainable speed. Here’s how, mathematically speaking: 

VO2max × SSmax / RE = fastest aerobic speed

Now the units there work out to km/min but it’s an easy unit conversion to pace, e.g. minutes per mile. When you look up your threshold pace in a chart (e.g. Daniels’ “T” pace), what you are seeing is this speed—which you might think of as vSSmax for “velocity at maximum metabolic steady state” (or vLT2, vMLSS, etc.) 

The origins of resilience

The idea of using the big three to determine a runner’s fastest aerobically-sustainable pace comes from a 1991 paper by Mayo Clinic physician Michael Joyner, which hypothesized that an athlete with extreme but plausible values for VO2max, maximum metabolic steady-state, and running economy could run 1:57 in the marathon. At the time, the world record was barely under 2:07, but it was Joyner’s model that eventually served as the inspiration for Nike’s Breaking2 project.

The 16 runners recruited for the Breaking2 project underwent rigorous in-lab testing of these three variables (with LT2 as the metric for estimating SSmax). Andy Jones, a British physiology researcher who conducted much of the testing on the Breaking2 athletes, published the (anonymized) data from these 16 world-class marathoners in 2021.  

Incredibly, the Breaking2 athletes had significantly better running economy and maximum metabolic steady-state values than the “highest possible values” used in Joyner’s 1991 paper. Dropping these values into the equation above gives a predicted marathon time of 1:55:05. 

Now, one of the athletes in the Breaking2 project happened to be Eliud Kipchoge, who did end up breaking 2:00 in the marathon, but even for him, 1:55 seems absurdly out of reach. 

What accounts for the five-minute gap between hypothetical and actual marathon performance? In other words, why can’t you sustain your fastest aerobically-sustainable pace for the entire marathon? 

Andy Jones interpreted this discrepancy as evidence for a missing “fourth dimension” of fitness that would explain the gap between theoretical and actual marathon performance.

In a paper published in August of last year, Jones argues that this fourth dimension of fitness is resilience—your ability to resist deterioration in the other three dimensions of fitness over a prolonged run (or a prolonged ride, for cyclists). 

What we know about resilience in running

The key idea of resilience is simple: two athletes with the same initial fitness level could have very different effective fitness levels at the end of a long workout or race. The resilience of the athlete determines how well they can maintain their initial fitness over the course of long bouts of exercise.

Cycling research has used a three-minute all-out sprint after a workout as a proxy for estimating SSmax, a method which has revealed two important findings.

First, resilience can vary enormously, even among athletes with similar initial maximum metabolic steady-states. The decrease in critical power (the cycling equivalent of critical speed, an estimate of SSmax) varied from practically zero to as 32% among a group of cyclists after two hours of riding in the high-end aerobic domain (i.e. above LT1, but below SSmax).

Second, resilience is an important differentiator between higher-level and lower-level athletes, even when those athletes have similar values for other physiological parameters. One study on professional cyclists showed that the main difference between WorldTeam (the top tier) and ProTeam (the second tier) cyclists was not their physiological parameters when fresh, but their physiological parameters after a fatiguing ride.

There’s less research in runners, but a quick look at performance data in the 10k, half marathon, and marathon reveals similar trends: some runners “convert up” much better to the marathon, suggesting they can sustain a very high percentage of their 10k or HM pace for the full marathon. Emile Cairess,  for example, has a relatively modest 10k PR by modern American standards (7 Americans have run faster than his PR just this year), but only three Americans ever have run faster than his marathon PR.

Since 10k pace and HM pace are heavily influenced by your (fresh) maximum metabolic steady-state—so much so that many people claim that HM pace is the same thing as lactate threshold pace—why is there so much scatter in marathon performance among runners who have similar 10k/HM fitness? 

The answer, in short, must be that these athletes differ when it comes to resilience. All of the evidence above strengthens the case that resilience is an independent dimension of fitness, separate from VO2max, SSmax, and running economy. 

The mechanistic basis of resilience in running

The cycling studies above establish that critical speed (or power, in cycling) declines after long, hard workouts, but they don't pinpoint the reason for the decline—it could be a loss of VO2max, a decrease in maximum metabolic steady-state, or a decline in economy. So, which of the big three declines in runners with poor resilience? 

Resilience and VO2max

A few studies have found that long, hard workouts cause a decrease in VO2max: stacking a progressive VO2max test at the end of a two-hour hard bike ride leads to a lower maximum value than doing a VO2max test fresh, for example. 

However, it’s not clear if this loss in VO2max comes “off the top,” while leaving maximum metabolic steady-state intact, or if your steady-state capabilities shrink in proportion to losses in VO2max. 

For example, suppose your fresh VO2max is 50 mL/kg/min, and your SSmax occurs at 80% of your VO2max (i.e. at 40). If your VO2max drops to 45 after running for two hours at marathon pace, what happens to your SSmax? Is it still 40, or does it “shrink” to 36? 

Losing VO2max off the top would not be so bad, since it would leave your SSmax alone. In contrast, if your whole range of aerobic power output shrinks proportionally, that’s much worse from a performance perspective. 

Resilience and steady-state max (SSmax)

It’s extremely hard to measure SSmax after fatiguing protocols, because the gold-standard metrics—MLSS and critical speed—require multiple in-lab visits to establish. Moreover, the raw lactate concentrations that you’d measure in a standard LT2 test are affected by changes in blood plasma volume (less blood plasma means a higher apparent lactate concentration) so it’s tough to do any reliable before/after studies.

Some work in cyclists using the aforementioned three-minute test does suggest that SSmax declines after long efforts, even when expressed as a rate of oxygen consumption. Research on runners, though, is scant.

Resilience and running economy

Running economy is the physiological factor where we can actually get some real traction when it comes to resilience. Because running economy is easy to measure, there’s a lot more research on how it changes during long, hard efforts. 

The best data on running economy in resilience come from this 1991 study on ten runners who completed a series of 15, 32, and 42 kilometers (an entire marathon), twice (!), on an indoor track (!!) with intermittent bouts of treadmill running to assess running economy. 

The study showed that the average runner experiences about a 5% decay in running economy throughout the length of the marathon. That’d mean that the same energetic expenditure that corresponded to 6:00/mi at the start of the race would yield only 6:20/mi by the end. This rate of decline in running economy was linear, suggesting that it’s the result of the slow, steady accumulation of fatigue over long distances.

5% was just the average increase in oxygen cost across individuals, though. What's more interesting is the variation from one runner to the next.

As with the more recent research in cyclists, the magnitude of variation is quite large—some runners show practically no change in running economy, even at the end of the marathon, while others may experience a 10% or greater increase in the energetic cost of running: enough to turn 6:00 miles into 6:40 miles!

The figure below shows the magnitude of this variability from a theoretical sample of 100 runners:

That variability in running economy is directly tied to resilience, and it’s not hard to see why: if the oxygen cost of running goes up later in the race, paces that used to be metabolically sustainable will now be unsustainable!

Why running economy gets worse in long races

Since declining running economy has the clearest link to resilience, it’s worth taking a look at some reasons why running economy gets worse over the course of a long race. Research supports three plausible mechanisms: muscle damage, localized glycogen depletion, and high-end aerobic fatigue.

Muscle damage

Long, continuous runs cause damage to muscle fibers, and that damage inhibits their ability to generate force. The result? Less efficient running. 

Muscle damage accumulates faster when running downhill—so much so that downhill treadmill running is used in scientific studies as a standardized way to cause muscle damage! Thirty minutes down a steep hill can drive up the oxygen cost of running by 4–7%, and coincides with a spike in blood biomarkers indicative of muscle damage. 

Incidentally, I suspect this effect is a major reason why many runners struggle at races like the Boston Marathon and California International Marathon—on paper, these courses have significant downhills, so they should be fast. But in both cases, the downhill is concentrated in the first half, so if you’re unprepared for the muscle damage induced by downhill running, your running economy will be significantly worse for the second half of the race. 

Localized glycogen depletion

Fast, high-end aerobic running relies heavily on carbs for energy, so glycogen depletion is a prime suspect for deteriorating running economy late in the marathon. 

Studies on whole-body glycogen availability are conflicting: one study on cyclists suggested that fueling with carbohydrates during a two-hour ride helps preserve critical power (the cycling equivalent of critical speed, and an estimate of SSmax). 

In contrast, a smaller study on nine runners found no effect of short-term carbohydrate depletion on running economy—though in that case, the 75-minute run may not have been a sufficient challenge to induce the critically low glycogen levels seen late in the marathon. 

One factor that may matter more than whole-body glycogen depletion, though, is localized glycogen depletion at specific locations in muscle fibers called triad junctions. Glycogen stored in these locations is used to pump calcium ions during muscle contractions, and when glycogen stored near these junctions is depleted, the muscle fiber can fail to respond to neural signals, even when the muscle as a whole still has glycogen left! 

Central fatigue

Running at speeds just below SSmax (e.g. running at half-marathon pace) results in a metabolic steady-state—blood lactate, oxygen consumption, intramuscular phosphocreatine, etc. are all stable, but even so, you can’t sustain half-marathon forever. You still get fatigued, even though the metabolic situation in your muscles is stable. Why?

One reason is localized glycogen depletion, discussed above, but another is central fatigue—a reduction in the ability of your brain to drive neural signals to your muscles. 

A study from 2007 investigated neural function in a region of the brain called the motor cortex (which controls muscle activation) following a treadmill marathon. After the marathon, motor cortex output dropped significantly, but only for neural signals sent to the leg muscles—while ankle strength decreased by 18% after the marathon, hand grip strength was unaffected. Research in cyclists using “single leg cycling”—which is exactly what it sounds like—finds similar results: motor cortex output is inhibited after a long ride, but only for neural signals sent to the “active” leg!

Collectively, these results suggest that central fatigue is a reaction in the brain to the overall biochemical situation in specific muscles after long bouts of high-end aerobic running. Physiologists have proposed various biomarkers that might be the cause of this fatigue (reactive oxygen species and calcium ions are two candidates), but in all likelihood, the brain probably uses a combination of different molecules and signals from “afferent” nerves, which send info to the brain about the local situation inside a given muscle.

Training to improve resilience

Understanding the mechanisms behind resilience can give us some clues on strategies for improving it. This is the part where we have to step out beyond the edge of what science supports and start drawing inferences ourselves. 

Given that the first authoritative review arguing that resilience is a distinct aspect of fitness was published less than a year ago, it almost goes without saying that there’s very little experimental work on how to improve resilience: we necessarily have to get out into more speculative territory.

Here are some workouts and strategies I have successfully used to improve resilience in marathoners and trail runners, based on my analysis of the biological roots of resilience. 

Training to resist muscle damage

Training to resist muscle damage is pretty straightforward: you need to do more mileage, longer long runs, and incorporate runs and workouts through rolling hills. 

More mileage and longer runs are more or less self-explanatory—the more you run (in a week, or in a day), the more accustomed your muscles will be to the repeated loading you’ll encounter in the marathon. Pace matters only to the extent that the faster you run, the greater the muscular forces you generate: from a muscle damage perspective, there shouldn’t be anything special about damage incurred at different speeds.

That's in contrast to other physiological adaptations, where certain intensities do have different effects (for example, above SSmax vs. below SSmax).

Runs through rolling hills are worth perhaps a bit more detail. As mentioned earlier, downhills are particularly damaging to running economy, so it can be worth leveraging this effect to improve resistance to muscle damage. On this front, a few sessions I like are:

• Long runs through steady rolling hills
• Repeats of 2–3 km at 8k–10k pace through rolling hills (e.g. 4 x 2km through rolling hills w/ 3min jog rest)
• Fast continuous runs or tempo runs that either start or finish with significant hills
• Hill repeats with “bookend” reps downhill

That last one is more advanced. Let’s start from a classic hill session: 10 x 2min at ~optimistic 5k effort, up a hill of ~5% grade with an easy jog down the hill for recovery. I have experimented with modifying this workout as follows: 

Starting from the top of the hill, do: 

• 1 rep fast downhill (covering the distance you’d cover in 2min going up)
• Easy jog down
• 4 x 2min fast uphill w/ easy jog recovery
• After the 4th uphill rep, 1min walk rest
• 1 rep fast down the hill again 
• Walk back to top of the hill, repeat once more for a total of 2 sets. 

So we’ve swapped in some fast downhill running on either end of each set to intentionally put a lot of stress on the muscles. The fast uphill running in between ramps up glycogen consumption in those muscles that just sustained some damage.

Again, advanced and ad-hoc, but seems to work after trying sessions like this with a few athletes. 

Training to resist glycogen depletion

Training to resist glycogen depletion requires, well, depleting your glycogen stores. Or fueling aggressively to delay depletion in the first place. 

Let’s talk about the latter option first, since it’s the easier one. On paper, it sounds like an easy fix: use more aggressive fueling. For marathoners, I usually start with 50–60 grams of carbs per hour, and experiment with higher fueling rates as tolerated. 

Online, I see people claiming to achieve fueling rates of 90–100+ g/hr, but I’ve only been able to get trail/ultra runners up to this level of carbohydrate intake. For most of my road marathoners, their stomach typically just can’t take it. Even 60 g/hr is a stretch for some of my athletes. If you do want to fuel aggressively, make sure you use a fuel source with both glucose and fructose.

Higher fueling rates are worth a shot if your resilience is poor, but realize that you may be “stomach-limited.” Remember also that it’s not necessarily whole-muscle glycogen depletion that’s the problem, but localized muscle depletion at triad junctions, something I suspect requires more specific training to target.

So, getting back to our first option: depleting your glycogen stores. Contrary to what most runners think, garden-variety long runs are insufficient on this front: the pace is too slow, so you don’t burn through your carb reserves. At a typical easy run pace, an average marathoner could go 30 miles or more before approaching glycogen depletion.

It’s worth mentioning high overall mileage again, because while individual easy or moderate runs don’t burn a ton of glycogen, the chronic effect of lots of easy to moderate running puts a steady drain on your intramuscular glycogen (which is part of why you feel fatigued when racking up lots of mileage, and why eating a high-carb diet tends to mitigate this high-mileage fatigue).

To really burn through glycogen, though, we need to run long and fast. Canova-style marathon training is great in this regard, since it makes heavy use of long fast runs at 90–95% MP, plus repeats at 100–105% MP with fast recovery at 85–95% MP. 

If that’s not enough, though, you can try using depletion workouts: long and fast workouts that are done with no breakfast beforehand, and no fuel during the workouts. Again, these are very advanced sessions and (in my opinion at least) are only appropriate for runners doing 65–70 mi/wk or more. 

That said, to aim to deplete glycogen, I have used progressive step-wise sessions like this: 

(Done in the morning, with no breakfast beforehand and no fuel during):

• 12 minutes progressing easy to strong + 4 strides
• 5 mi at 92% MP
• 5mi at 93–94% MP
• 5mi at 95% MP
• 3–5mi at 96% MP progressing to 98% MP, if possible
• (No cooldown)

Just as a warning, this is a very tough workout. You would only do a tough depletion session like this maybe two or three times in a marathon buildup.

The progressive structure of this long fast run helps really ramp up the need for glycogen late in the run. Sometimes you may fail to complete the final chunk, which is fine. In my experience, you don’t need to completely crash and burn to benefit from a depletion workout, but if it happens, it’s not a bad thing. 

Some people think the purpose of this kind of session is to force your body to oxidize fat. I disagree. I think the purpose is to encourage your body to store more glycogen, especially at key locations like triad junctions.

I have also used classic MP workouts, e.g. 5 x (4k at 100% MP, 1k at 90% MP) done in a depleted state, again giving the athlete the option to push the last repeat faster if possible to really “flush out the tank.” That’s still a challenging session, but not quite as demanding as the stepwise long fast run above. 

A lighter, introductory option for lower-mileage runners might be something like 12mi easy + 5–7mi progressive, again done depleted and again aiming to finish basically as fast as possible. 

Training to resist central fatigue

Training to resist central fatigue is, again, relatively simple: we need to spend a lot of time running in a state of moderate to high central fatigue. The way to do this is with high-end aerobic workouts: roughly, sessions at 80–92% 5k pace, or 90–105% MP.

When it comes to high-end aerobic workouts, you’ve got several options on this front:

• Daniels-style cruise intervals and “T” pace tempos (~92% 5k)
• Norwegian-style double threshold sessions
• Long fast runs at 80% 5k pace
• Fast continuous runs at 85% 5k pace
• Fast continuous runs at 90% 5k pace
• Kenyan-style progression runs
• MP and HMP-specific workouts

In all cases more volume is probably more important than going faster, at least from the perspective of improving resilience. Do remember, though, that workouts in the high-end aerobic domain do much more for your overall fitness than just improving resilience!

It’s worth noting that for resisting central fatigue, high-end aerobic workouts are strictly superior to racking up lots of easy mileage, because easy and moderate running do not generate the right kind of fatigue. 

To be sure, you do get tired if you do a 20 mile long easy run, but that is a very different kind of “tired”, at a biochemical and neurological level, than the fatigue you encounter during, say, 10mi at 85% of 5k pace, or 7 sets of (3k at 100% MP, 1k at 90% MP). 

In a more technical sense, easy and moderate running are in a different domain of intensity than high-end aerobic running. To improve resilience on this front, we need to do workouts in the high-end aerobic domain, not the easy to moderate domain.

Some shortcomings of resilience theory

I’d be remiss not to mention some of the shortcomings to the current picture of resilience as a dimension of fitness. The biggest one is the fact that the current theory doesn’t explain why running even splits seems to be the optimal way to run a marathon—if your fitness is deteriorating throughout the race, why not run fast early, then “ride the wave” as your SSmax gradually fades? 

Many marathoners do in fact adopt this kind of pacing strategy, but it’s clearly not the most effective: all world and national records in the marathon are set using even or slightly negative splits, not a steady slowdown.

It might be that there’s a nonlinear process at work—for example, maybe your running economy deteriorates at a faster rate when you run closer to SSmax. An effect like this would make “fast now, slow later” strategies worse overall than an evenly-paced race. 

It’s also not entirely clear how to quantify resilience numerically. Deterioration in running economy is simple and easy to measure in the lab, but it won’t capture decreases in VO2max or SSmax. I expect a raft of research in this direction over the next five to ten years. 

Right now, I think of resilience as the fraction of SSmax that you can sustain for the marathon. This way of thinking nicely integrates into the “Big Three” framework, in that marathon pace is now just: 

VO2max × SSmax × Resilience / RE

But that’s a very marathon-centric way of thinking about things, and doesn’t offer much to ultramarathoners (to say nothing of cyclists).

Does deteriorating fitness late in the race mean you must slow down?

One last thing to remember regarding resilience—just because your economy or VO2max or SSmax is deteriorating doesn’t spell doom in the marathon.

You do have one ace in the hole: your anaerobic energy reserves. These reserves are finite, but just like in a 5k or 10k, you can gradually spend down this energy over the last several miles of a race to counteract the mounting metabolic cost of running.

Here’s a plot of heart rate and pace data throughout a marathon that shows this principle at work:

Notice how the pace stays the same in the final few miles of the race, even when heart rate (and presumably the metabolic cost of running) starts to rise. That difference is being made up by this athlete’s anaerobic energy reserves.

Now, remember that heart rate is far from a perfect stand-in for metabolic cost, but even so, you can see the dramatic rise in heart rate over the last several miles of the race, contrasted with the stable pacing. This runner’s metabolic rate in the final few miles of the race is almost surely unsustainable, but that’s ok—they only need to last a few more miles.

In contrast, if this deterioration in running economy happens at 16 miles instead of 22 or 23 miles, you’re in real trouble. Pushing back the onset of that deterioration is the task of resilience-focused training. 

Recap

Physiological resilience measures how well you can maintain your initial level of fitness—as quantified by your VO2max, maximum metabolic steady-state, and running economy—throughout a long effort like a marathon or trail race. 

Runners with high resilience can maintain the same pace for a long time, while runners with poor resilience tend to fall apart in the latter stages of a marathon or ultramarathon. 

Most of what we know about resilience so far is about its effects on running economy. Long, fast runs induce a steady loss of running economy, though this deterioration in economy varies widely from person to person. Runners with poor resilience might see a 10% increase in the metabolic cost of running by the end of a marathon. 

Deteriorations in running economy are likely the result of a combination of glycogen depletion at key locations within muscle fibers, damage to muscles induced by eccentric contractions, and central fatigue in the spine and brain. 

The best strategies for improving resilience are high mileage, workouts through hilly terrain, and long runs and intervals at 90–105% of marathon pace. Advanced runners could consider “depletion workouts,” i.e. doing long, fast training sessions without taking carbohydrates before or during.

Resilience offers a lens into marathon and ultramarathon performance beyond VO2max, maximum metabolic steady-state (SSmax), and running economy. If you find yourself falling apart in the later stages of long races, even when the pace early on feels manageable, training to improve your resilience could be the solution.

Learn more about the science of training

If you enjoyed this article, subscribe to my newsletter! It’s the best way to find out when I’ve got new content coming out. In fact, my newsletter subscribers got early access to this article on resilience and had the chance to read it over a month before it was posted here on my website! Sign up if you want to be the first to see what I'm working on.

I also have a book, Modern Training and Physiology for Middle and Long-Distance Runners, that focuses on the science of performance and training for events from 800m to the 10k. Check it out!

Related articles