Training program for UTMB, my 2017 apex race || Interview with David Roche

How have I trained for Ultra Trail du Mont Blanc (UTMB), the world’s premier ultra marathon? As part of an ongoing series, I interviewed again my coach, David Roche, transcript below.

But to fully understand the context of this training, it may be necessary to read similar interviews prior to my two other big races this year:

Boston and Bighorn were worthy in their own right, but they were critical stepping stones to UTMB. My training for Boston resulted in a dramatically better running economy, or the effort involved in running at various paces. And during Bighorn training I saw a huge increase in my vamAT, or the amount of vertical climbing I could do (per time) at my aerobic threshold.

If I stuck to my old ultra training program (“volume, volume, volume”), I would finish UTMB, but would run/hike slower at every energy output level.


Q | We divided the 11 weeks between Bighorn 100 and UTMB into several shorter training phases. What were these phases and what was the objective in each?

After Bighorn, we had more time to do a deliberate build for a mountain ultra than we did between Boston and Bighorn. There were three special considerations in the training strategy. You:

  • Carried your best climbing fitness into Bighorn, and that fitness didn’t need to be rebuilt.
  • Recovered relatively quickly from Bighorn because of the sh*t-soup trail conditions in the second half. And,
  • Had scheduled the Pfiffner Traverse Yo-yo, a 160-mile 9-day backpacking trip during the peak of the build phase for UTMB.

In terms of timing, we had a few weeks, then the Pfiffner, then a few weeks, then taper, providing a natural dividing line for the phases.

After a week of recovery, we did a brief running economy rebuild, providing a neuromuscular stimulus to lay the groundwork for efficient mountain running. This consisted of strides and short intervals with equal to double rest. You responded better than expected, even running some blazing fast 200s. Your vVO2 was comparable to pre-Boston.

Usually, when short-distance running economy is looking that good, we’d jump into a lactate threshold or hill phase, like before Boston and Bighorn. For example, I’d have you do 5 x 3-minute hill repeats.

But given your hill fitness and the upcoming Pfiffner, we blended the LT/hill/ultra phases beginning a week before the Pfiffner. The hill efforts within longer runs reinforced your vamLT and vamAT, which was already where it needed to be for UTMB.

The traverse provided a big aerobic stimulus, which is going to be important during the 106-mile adventure in the Alps. So then we jumped into the true bread-and-butter of ultra performance: the pure ultra phase. Three times a week we’d go into the mountains, with varying focuses each day, designed to make a training lasagna. First layer was the meat: 100 mile effort, honing the aerobic system. Second layer was the pasta: mixed effort runs focused on uphill or downhills, dialing in vamAT and resilience. The final layer was the delicious cheese: moderate or hard efforts that bring it all together, culminating with the Pawnee Buchanan FKT.

Now, it’s taper, then some lasagna feasting next week!

Q | The Pfiffner Traverse Yo-Yo was a 9-day 160-mile backpacking trip with 60k vertical feet of climbing at elevations of 10,000-12,000 feet. Was it helpful to my UTMB training, or would I have gained more by a more usual training program? If a runner wanted to do something similar, do have any recommendations on how or when to integrate it?

Hearing the details of that traverse again makes me want to take a sympathy nap.

Given your unique background, it was helpful. It reinforces three things that will be key on race day:

  • Physical resilience to all-day pounding,
  • Hiking ability, and
  • Mental toughness.

On the flip side, it temporarily reduces running economy and for most people, that particular adventure would come at too great a recovery cost.

Most athletes planning a 100-mile race could benefit from long backpacks so long as they manage the downsides. Just like a 50-mile or 100-KM race is optimal 5 to 10 weeks before a 100-miler, this type of adventure could replace a long ultra (or supplement a 50k) in this time frame. But you must approach the adventure as if it’s a race, i.e. physically recovered beforehand and physically recovered afterwards before an economy rebuild.

On average, the Pfiffner Traverse climbs or descends 800 vertical feet per mile. One day, I climbed 6,000 vertical feet in just 13 miles, an average of 925 vertical feet per mile.

Q | Let’s return to the phases. What’s a normal recovery period after a 100-miler, assuming a well-trained athlete and a max effort? And, if you start the next training cycle before a sufficient recovery, what would you expect to observe?

Here is where the ultra version of the dose-response curve comes in. Stress and recovery are non-linear, depending on a multitude of individual factors. But for the purposes of simplifying, a full-throttle 100 takes around a full month to recover from for most athletes, generally segmented as:

  • Week of rest,
  • Week of slow running,
  • Week of slow running plus a couple short hills, and finally a
  • Week of easy running.

Only after this full recovery is the athlete is ready to ease back to somewhat normal training. For example, after winning Western States on June 24, SWAP athlete Cat Bradley didn’t do her first higher heart rate run until the beginning of August.

This is not an idle question — the stakes are high in 100 recovery. When an athlete returns too soon, it can manifest similarly to overtraining syndrome. You’ll see a week or two of solid running, followed by increased lethargy, decreased performance, and sometimes illness (like a persistent cough, or even mononucleosis-like symptoms). It’s especially risky for athletes doing their first 100. Meanwhile, for someone as experienced as you, we can go a lot more on feel and trust.

Only then is the athlete is ready to ease back to somewhat normal training, with full fitness and recovery around 2 months.

Q | We addressed running economy extensively in our first interview, and I don’t feel like we need to go there again. If a reader is not familiar with the concept, I would strongly encourage them to read that interview. My performance gains this year are probably due more to my massively improved running economy than to any other factor. David, anything to add to that?

Running economy is a catch-all term that incorporates lots of factors, from biomechanical (i.e. muscle tension) to neuromuscular (form efficiency) to cardiovascular (cardiac stroke output). By improving your economy early on and then re-introducing economy phases throughout the year, every run is just a bit better, allowing us to more effectively target the energy systems that count on race day (here, LT and aerobic threshold).

Essentially, a rising running economy tide raises all fitness ships, whether an athlete is racing a road marathon or trail ultra (or even doing something seemingly distant from our emphases, like fast-packs).

Q | During the 3-week “pure ultra phase,” most long runs had a specific emphasis, beyond just volume. Examples: a hard 25-minute uphill effort near the start of the run, or 8 x 3-min VO2 hills at the end, or a UTMB effort from start to finish (low-130’s average HR), or a FKT attempt. How valuable was this specificity, relative to pure vanilla long runs?

Vanilla is a great way to describe long runs without specific emphases. It’ll probably get the job done, but double chocolate chunk is way better.

It gets back to the component parts of reaching your potential at ultras. General aerobic fitness and resilience are two main components, and the vanilla long run works those elements well. But add in some spice, and we can additionally target other components, like vamLT (hill climb at start) or vamAT (FKT attempt) or metabolic efficiency (UTMB effort). Then the components all build upon each other over the course of a training cycle.

It’s as if you’re trying to build a stool to reach an apple that is just out of reach. You could just get a bunch of wood, then throw it on the ground, and you’ll probably get there. But if you carefully stack the wood, it’ll be more stable, and you probably will need less wood in the first place.

Q | Finally, the taper. My last long run was two weeks prior, and otherwise it’s all easy runs with single-digit or low-teens mileage. Why don’t these days include “sharpening” workouts, like they did prior to Boston?

It’s the old inverted pyramind of training — start more general, get specific as we go. We’re now at the pointy end of the pyramid, the sharpening phase. For the marathon, being sharp means optimizing your velocity at Lactate Threshold (vLT), with some velocity at aerobic threshold (vAT) mixed in.

For UTMB, where you’ll be on-course for about 24 hours, specific training means low-level aerobic efficiency, working on slow twitch muscle fiber recruitment and metabolic efficiency.

Most of the runs are more relaxed because we want all of your body’s energy channeled into the systems it will use on race day. And they are shorter to avoid flooding your body with stress hormones prior to a stressful travel-and-race environment.

If you were doing an ultra like Western States, which would be far faster and slightly more intense, we’d be doing some hill climb tempo work. If UTMB were in the Front Range of CO where you live, we’d be doing a few extra miles due to diminished stress. But given the unique demands of the race and the location, the pointy end of this inverted pyramid means that sharpening is dull: lots of easy, efficient runs.

Q | Let’s finish with a few questions about race day. What’s the optimal pacing for a hard course like UTMB? Is it realistic to run even splits with an increasing effort, as I did at the Boston Marathon? Or should I hope to simply slow down less than everyone else?

Negative splitting UTMB would be witchcrat, or would at least require us to throw you in the river to see if you float . In a mountain 100 like UTMB, effort can’t be viewed in a linear way. Down periods are impossible to avoid completely, and power at aerobic efforts will decrease by 10-20 percent for most top athletes even assuming perfect pacing. With that in mind, the goal is to manage the down periods and the inevitable fade in power at aerobic effort.

We can manage that by imposing a strict cap on heart rate 5 beats below aerobic threshold on the uphills, which will ensure that we don’t bankrupt your legs prematurely by making a LT withdrawl.

On the downhills, in the first half, we just focus on relaxed, short strides, and not braking (but not pushing either), which will be lower heart rate still. In the 2nd half, the same HR cap rule applies on uphills, which will make them a bit slower. But on downhills, you can open it up a bit more. So our version of even-splits is slightly stronger downs to counteract slightly weaker ups.

Q | Final thing: Help me set expectations. Where should I expect to be in the field at Mi 25, Mi 50, Mi 75, and of the course the finish?

I am guessing you don’t want the answer: “purposefully smiling, purposefully smiling, reluctantly smiling, smiling as it rains confetti and the national anthem plays?”

This is the most stacked field in the history of ultrarunning. So a lot of what happens depends on factors outside of your control. But, assuming people run mostly in line with past races and you race to my expectations, you’ll be anywhere from:

  • 30 to 60 at mile 25,
  • 20 to 50 at mile 50,
  • 8 to 25 at mile 75, and
  • 4 to 15 at the finish.

I said when we first met that I expected top-5, and you did everything to make that possible. However, I didn’t account for the competition being so astounding, so if race day doesn’t go according to plan, it is my fault, not yours.

Unlike before Boston or Bighorn, I will actually make a prediction this time: 5th, moving up from the field and closing fast. But, in a truly touchy feely coach/friend sentiment, I am just proud of you for everything you have done over the last nine months. Whether you finish 5th or 50th or DNF, I’ll always look back on that time as a period when I grew as a coach and person because I got to be a small part of your life.

Hanging out with BFF Oden after yet another long run. My legs tell the story of the day and the season.

Posted in on August 27, 2017
Tags: , ,


  1. Steve Sims on August 27, 2017 at 6:49 am

    Look at that vascularity! At age 35 no less. GO get em bro. I am proud to have been your coach and friend for half your life. The man is ready. Do it smart and pick ’em off the last 20! Thanks Dave Roche for fine tuning this lion of a human being.

  2. Brent Warner on August 27, 2017 at 8:02 am

    Go get ’em, man.

    It’s been seriously fun following your training these past few months, and a few of your most recent workouts are absolute mind-blowing monsters. Cannot wait to see what you’re capable of on race day. Good luck!!

  3. John Wright on August 27, 2017 at 7:28 pm

    Hilary and I just read that out loud. That gave us chills at times and the game plan between you 2 is damn exciting and seems super effective. We can’t wait to cheer for you Friday!!

  4. Hunter Hall on August 28, 2017 at 12:58 am


  5. Scott on August 28, 2017 at 7:14 pm

    It seems you’ve asked about your expected placement at UTMB before, possibly as a starting point in your training.
    Do you really feel it’s an appropriate and manageable statistic? Doesn’t it rely on factors completely out of your hands far more than anything you can influence yourself?
    I personally feel that time goals in ultras, especially mountainous ones. are more centered on me (you) and are expressive of my training and execution than placement.

    • Andrew Skurka on August 29, 2017 at 9:45 am

      If the course is the same year to year, yes, I agree with you. It’s more relevant to say, “I think I can run 24 hours,” than, “I think I can be top 5,” since the latter goal might prove completely out of reach if the field becomes stacked (as it did this year). As another example, my goal at Boston in April was to run 2:32, and I didn’t even know before the race whether that’d be good enough for 50th, 150th, or 500th.

      The only value in a placement goal is if the course changes, like due to a re-route or exceptional weather. In that case, historical splits and finishing times have less relevance, and you’re merely fighting for place.

      Between you and me, I think David is being a little optimistic with Top 5. I’d have to run the race of my life for that, and a lot of other guys would have to implode. I’ll be happy with Top 10 and ecstatic with Top 10. If I’m Top 5, you’ll be able to hear me whooping from the across the Atlantic Ocean.

  6. Didier Deflem on August 29, 2017 at 8:56 am

    Hi Andrew,

    Interesting as always!

    Can you give us an overview of the mandatory UTMB gear you will be using?

    As the predictions, 4 days before the start, are pretty bad, proper gear selection could become a pretty important choice!

  7. Scott C on September 4, 2017 at 10:53 am

    Awesome post again, Andrew – thanks!

    What is your aerobic threshold? I remember your MAF was close to mine (~150), so was curious where the “cap” that David recommended was placed. (“We can manage that by imposing a strict cap on heart rate 5 beats below aerobic threshold on the uphills,”)

    Also, you’ve primarily blogged about running/hiking, can you or David share insight into core and leg strengthening for long days in the mountains? Would love to learn more about the strength component, though I know your hiking background probably helps significantly.

    Thanks again for all your blog posts, especially with David – they are a great resource.

    Great job at UTMB!

    • Andrew Skurka on September 6, 2017 at 1:59 pm

      I tried to keep my HR below 135 bpm from the very start. I think it only crept up to 141, and quickly I brought it back down. My avg HR was 121 through the race, which was lower than we expected it would be. I was thinking high-120’s, since I’d averaged 127 at Run Rabbit in 2015 and since I was at 119 and 123 for Silverheels 100 and Bighorn 100, respectively, and that had been sub-max efforts.

Leave a Comment