Durability is one of the defining qualities of a successful ultrarunner.
Broadly, we think of durability as the ability to keep running well deep into the race.
By that definition, Caleb Olson produced one of the clearest examples of durability at UTMB this year.
At Les Tseppes, around 146 km into the race, Caleb was almost 16 minutes behind Baptiste Chassagne. He crossed the finish line just 46 seconds behind.
If we were looking for the athlete who demonstrated the strongest ability to perform late in the race, Caleb would be an obvious candidate.
But he didn’t win.
Ben Dhiman did.
Ben opened up a huge lead early and crossed the finish line almost 30 minutes clear.
And the women’s race gives us another problem. Blandine L’Hirondel arrived at Vallorcine, around 158 km into the race, almost 46 minutes ahead of Rachel Entrekin. She slowed considerably over the final 18 km and gave back roughly 19 minutes, yet still won by almost 27 minutes.
So if durability is about how well an athlete performs as fatigue accumulates, why didn’t the athlete who appeared to demonstrate the most durability win?
There are already some useful ways of thinking about durability, and I don’t disagree with the concept itself. But where things become more complicated is when we use late-race slowdown as the main way of explaining performance over 100 miles.
The athlete who slows down the least isn’t necessarily the athlete who wins, and the training needed to improve an athlete’s ability to perform well late in a race isn’t necessarily the same for everyone.
The splits from the top three athletes illustrate just how much more nuanced the relationship between durability and performance in a 100-mile race can be. All three finished on the podium, but they got there in very different ways.
To see exactly how different, I went back through the official UTMB split data and looked at where they gained and lost time across different stages of the race and different types of terrain.
I looked at both the change in the overall gap between athletes and their moving time over each section. Those aren’t always the same, because the overall gap also includes time spent at checkpoints and aid stations.
I then grouped the course into three broad terrain categories:
* climbing
* descending
* rolling/mixed
This lets us see where each athlete gained and lost time, and how that changed as the race went on.
So where did Ben win the race?
His advantage over Baptiste was already substantial by Courmayeur.
| Timing point | Ben’s lead over Baptiste |
|---|---|
| Courmayeur | 16:37 |
| La Fouly | 23:30 |
| Champex-Lac | 26:56 |
| Vallorcine | 29:33 |
| La Flégère | 30:12 |
| Finish | 29:56 |
Ben built a race-winning advantage early, extended it through to Vallorcine, and then maintained it all the way to the finish.
The lead changed how he raced
But there is an important tactical element here.
Ben had a huge lead, but he still had to decide how hard he needed to run to protect it.
In his interview afterwards, he described finding himself looking down the barrel of a “90-kilometre time trial” once the race had opened up. He knew that wasn’t necessarily the ideal way to run a race, but he was able to use his knowledge of the course, how he was feeling and the information he was getting about the gap behind to decide where he could push.
He talked specifically about running the descents hard and choosing where he could gain more time.
So although the split data shows us that Ben maintained his advantage, his own account gives us some insight into how he managed it. He wasn’t simply following a predetermined pacing plan. He was continually adjusting his effort based on how he was feeling, the terrain and the gap behind.
Once the gap was established, the objective wasn’t to find his absolute fastest possible time to Chamonix.
It was to “hold the gap and win the race”.
That matters when we interpret his late-race performance.
Caleb gives us a very different picture
Caleb Olson is probably the most useful athlete in the race for understanding how durability can look very different from one athlete to another.
He finished third, but spent much of the second half chasing Baptiste.
At the major checkpoints, Caleb’s deficit was:
| Timing point | Caleb behind Baptiste |
|---|---|
| Courmayeur | 5:06.83 |
| La Fouly | 10:07.28 |
| Champex-Lac | 11:52.26 |
| La Giète | 13:17.00 |
| Trient | 13:27.56 |
| Les Tseppes | 15:58.00 |
| Vallorcine | 11:23.98 |
| La Flégère | 6:32.00 |
| Finish | 0:45.91 |
At Les Tseppes, around 146 km into the race, Caleb was almost 16 minutes behind Baptiste.
From there, he started taking the time back.
By Vallorcine he had removed more than four minutes. By La Flégère he had removed another five minutes, and over the final descent he removed another five minutes and 46 seconds.
He crossed the finish line only 46 seconds behind Baptiste.
The split data shows a clear change in the second half, and Caleb’s own comments give us some insight into how he had approached the race.
At La Fouly, with roughly 40 miles still to run, he said he felt good about the second half and had “saved some energy for it”. Earlier in the race, he had deliberately backed off when he felt the pace was too hard, choosing not to follow every move at the front.
This fits well with the splits: Caleb had been losing time to Baptiste through the middle of the race, but late in the race he started taking that time back.
So where was Caleb gaining that time?
In the climbs from Champex to La Giète and Trient to Les Tseppes, Caleb was 5 minutes 39 seconds slower than Baptiste.
But on the descents from La Giète to Trient and Les Tseppes to Vallorcine, he was 4 minutes 23 seconds faster.
Then something changes. On the climb from Vallorcine to La Flégère, Caleb was 5 minutes 16 seconds faster than Baptiste.
On the final descent to Chamonix, he was another 5 minutes 46 seconds faster.
Earlier in the race, Caleb wasn’t consistently outperforming Baptiste on the descents. The important change was that late in the race he was able to run faster than Baptiste both uphill and downhill.
Earlier, Baptiste was the faster athlete. Later, Caleb was the faster athlete.
Caleb had deliberately held something back early, and that allowed him to finish much stronger than Baptiste.
The split data shows that Ben built a huge lead and maintained it to the finish. But that doesn’t necessarily tell us how durable he was. Once he had that margin, he didn’t need to find out how fast he could still run late in the race — he needed to run fast enough to protect the lead.
Then there is Blandine
The women’s race gives us a different example of how durability can contribute to a win.
Blandine L’Hirondel built a huge advantage over Rachel Entrekin, and that advantage continued to grow through the middle of the race.
Her lead was:
| Timing point | Blandine’s lead |
|---|---|
| Courmayeur | 6:12 |
| La Fouly | 25:53 |
| Champex-Lac | 38:19 |
| Vallorcine | 45:55 |
| La Flégère | 36:57 |
| Finish | 26:57 |
By Vallorcine, around 158 km into the race, Blandine was almost 46 minutes ahead.
Over the final 18 km, she gave back roughly 19 minutes. The race footage makes it clear that those final kilometres were becoming much harder for her. Her descending performance dropped considerably, and she had two significant falls on the way to Chamonix.
Although Blandine’s descending performance visibly deteriorated she kept the lead and crossed the finish line almost 27 minutes ahead of Rachel.
This looks quite different to Ben’s race. Ben built a large advantage and was able to maintain it almost all the way to the finish. Blandine built an even bigger advantage, then gave back almost half of it over the final 18 km.
The difference appears to have been partly deliberate. In the COROS analysis of her race, Blandine described deliberately holding back early and then targeting the middle of the race, from roughly 30 to 140 km, as the part where she wanted to push harder.
Where did Blandine gain the time?
When we look at where she built that advantage, climbing wasn’t the dominant factor.
From the start to Vallorcine, her advantage over Rachel was:
Climbing: 3:51
Descending: 22:24
Rolling/mixed: 18:01
The biggest gains came from descending and mixed terrain.
In simple terms, Blandine made hay while the sun was shining. She used the parts of the course where she was strongest to build an enormous lead, and that gave her enough margin to absorb a significant slowdown over the final 18 km.
She didn’t need to be the strongest athlete over the final 18 km.
She had already done enough earlier in the race that she didn’t need to be.
What does durability actually mean?
We tend to think of durability in terms of what happens late in the race. The athlete who is still running well when others are slowing down is the one we describe as durable.
But that doesn’t necessarily tell us why they performed better over the whole 100 miles, and the athletes we’ve just looked at show why it isn’t quite that simple.
Late-race performance is the result of more than one thing.
How an athlete distributes their effort through the race matters.
So does how well they are able to keep performing as fatigue accumulates.
These are related, but they aren’t the same thing.
An athlete can deliberately hold something back early and still struggle when the race gets long. Conversely, an athlete can slow considerably late in the race and still have run an excellent 100-mile race if they have created enough of a margin earlier.
This is why simply asking “Who was the most durable?” misses an important part of the picture.
We need to look at what the athlete did earlier in the race, what they had left when fatigue really started to bite, and what the race required of them at that point.
What does this mean for training?
This is where I think the lesson becomes much more useful than simply saying:
Train durability.
Of course we want athletes who can perform deep into a 100-mile race.
But the more useful coaching question is:
What does this athlete need to be capable of doing for 100 miles?
Two athletes can have exactly the same goal and need completely different preparation.
This is where I think the usual advice around training durability falls short.
The recommendations are generally broad: run more, do more downhill, do back-to-back long runs, spend more time on your feet. All of these can improve an athlete’s ability to perform late in a race.
But they don’t tell us why an athlete is slowing down, or what that particular athlete needs to improve.
If we want to make an athlete better over 100 miles, we need to identify what is actually limiting their performance and train that.
There isn’t one type of training that makes every athlete more durable.
The training needs to match the athlete.
Ben provides a good example.
Before UTMB, he identified climbing as an area where he needed to improve and deliberately increased his climbing volume by around 30%. Afterwards, he said he was faster across the board. His broader point was that training has nuance: there are broad principles, but the work still has to address the individual athlete.
Caleb gives us another example, but with an important qualification.
He said he likes the variety of terrain at UTMB and was trying to “shape my training and shape myself into the kind of person that this suits.” He therefore spent a lot of time looking for terrain in the Wasatch that would replicate the demands of the race.
That makes sense for Caleb.
But it doesn’t automatically mean that every athlete trying to improve at UTMB should spend more time on UTMB-like terrain.
UTMB has a lot of climbing, hiking and descending, so the obvious conclusion is to train more climbing, hiking and descending.
But it depends on the athlete.
One athlete might genuinely need more hiking and descending because those are limiting their performance.
Another might already be very good at both, but lack the aerobic fitness or running speed to move quickly on the more runnable sections. Giving them more hiking and descending isn’t going to solve that problem.
The demands of the race tell us what the athlete needs to be able to do on race day. They don’t necessarily tell us what the athlete needs to train more of.
That’s the job of the coach: to identify what is actually limiting the athlete’s performance.
Not simply:
“How do I make this athlete more durable?”
But:
“What is stopping this athlete from running to their full potential over 100 miles?”
That distinction plays out on the racecourse as much as it does in training.
How an athlete runs a 100-mile race is influenced by their own strengths and weaknesses, as well as the demands of the course.
And if you’re racing for position rather than simply against the clock, there is another layer again. You’re making decisions about when to push, when to hold back and how much risk to take based on the race around you.
That’s what made this year’s UTMB so interesting. The athletes at the front raced the course in very different ways.
For the rest of us, the important question is what is actually limiting our ability to run our best 100 miles — and how do we train to improve it?
I’ll explore that in more detail in a future article.
