

About the Author
Albertus Roux is a sports scientist and ultra-trail runner specialising in endurance performance and human physiology. Drawing on scientific research, coaching experience, and years of ultra-distance mountain running, his work focuses on resilience, discipline, and the mental and physical demands of endurance sport. He helps athletes improve performance and unlock long-term potential through evidence-based insight and real-world experience.
Part 3: From data to training: applying this in the real world
This is the final part of a 3-part series based on my Honours dissertation:
A multivariate model predicting ultra-trail running performance.
In Part 1, we explored why ultra-trail performance is so difficult to predict and why traditional endurance models fall short.
In Part 2, we looked at what the data actually revealed from a group of 19 runners competing in the 2024 Ultra Trail Cape Town 100 km.
Now we arrive at the most important question:
What does this mean for how we actually train to become better trail runners?
Because research is just noise unless it changes behaviour. And in a sport as complex and unforgiving as ultra-trail running, behaviour is everything.
The core message (If you read nothing else)
Ultra-trail running performance is not the result of physiology alone.
It emerges from the interaction between:
- Physiological capacity
- Long-term training exposure
- Terrain- and environment-specific stress
- Real-world performance behaviour
Your engine matters.
But what matters more is how well you can keep using that engine when conditions are far from ideal.
How this research changed my coaching
For a long time, I evaluated athletes primarily through physiology.
How big is your engine?
What’s your VO2max?
Where is your ceiling?
How fast can we move the numbers?
In predictable sports like road running, this approach makes sense.
This research showed me that ultra-trail running is one of the clear exceptions.
The strongest predictors of performance were not lab-based metrics, but how well athletes had adapted to the specific demands of trail running over time. A solid physiological base is necessary to tolerate training, but it is the training itself that ultimately matters more.
Now, when I meet a new athlete, the first questions I ask are very different:
- What surfaces have you trained on most?
- How many years have you trained consistently?
- How much elevation gain do you accumulate per month and per year?
- What does your trail racing history look like?
These questions tell me far more about future ultra-trail performance than a VO2max number ever could. They guide training decisions in a way that reflects the reality of the sport.
The problem with chasing perfect numbers
It’s easy to fall in love with numbers. They feel precise, objective, and reassuring. Lab testing and clean graphs going up can boost confidence, and that confidence does matter.
Traditional lab testing still has value. It helps identify athlete profiles and can support objective decision-making. But precision does not equal relevance.
In this study, traditional endurance metrics such as:
- VO2max
- Lactate threshold
- Running economy
were consistently outperformed by simple, real-world variables:
- Total elevation gain over the previous 12 months
- 5 km personal best
- Distance-to-elevation ratio
The implication is clear.
An athlete with average physiology but years of trail-specific exposure will reliably outperform an athlete with world-class physiology but poor specificity. This may sting if you are naturally gifted. It is also deeply empowering if you are not, because it shifts the focus away from genetics, which you cannot change, and toward training behaviours, which you can.
A real-world training hierarchy (according to the data)
Based on the findings, the variables that mattered most for ultra-trail performance followed this general hierarchy:
- Consistency over years
- Total elevation gain (and loss)
- Simple performance markers (5 km PB, marathon PB, trail benchmarks)
- Physiological capacity
Side note: I strongly suspect that the percentage of training time spent on actual trail versus road belongs very high on this list as well. We simply did not have that variable available in the dataset.
This hierarchy should fundamentally change how we coach and how we assess progress.
Spend less time asking:
- How fit is this athlete?
- How fit do they need to become?
And more time asking:
- How consistent has their training actually been?
- What terrain are they training on?
- How much elevation are they accumulating?
- How is performance trending over time?
Designing training around reality
Ultra-trail running is chaotic by nature.
Terrain changes.
Weather turns.
Fatigue shows up unpredictably.
Gear fails.
Nutrition plans fall apart.
Yet many training plans are still built around tidy weekly structures and idealised sessions. This research suggests that what matters most is not perfect execution, but repeated, sustainable exposure to the stressors you will face on race day.
Most of us are normal humans with jobs, responsibilities, and limited time. That means training must be intentional, not perfect.
Prioritise:
- Long, sustained climbs
- Prolonged descents under fatigue
- Variable pacing strategies
- Technical and uneven terrain
- Training in adverse conditions
- Practising nutrition and hydration under stress
Not occasionally. Systematically.
Designing training around reality does not mean abandoning structure. It means building structure around the environment you will race in. Your training may not be perfectly optimised for improving physiology alone, and that is okay. You are training your ability to keep using your engine when conditions are far from ideal.
Closing thoughts
Ultra-trail performance is not built in a single training block or explained by a single variable.
It is constructed slowly through thousands of small, unglamorous exposures that teach your body how to perform despite the circumstances.
This is the heart of the sport.
It’s not what you can do when everything goes right.
It’s what you can still do when everything goes wrong.
FAQs
How should I apply ultra-trail running research to my training?
Focus less on lab-based numbers and more on long-term consistency, elevation gain, terrain exposure, and simple performance markers that reflect real-world trail demands.
Is physiology still important for ultra-trail running?
Yes, physiology provides the foundation, but it explains far less performance variation than in road running. Training specificity and exposure matter more in long, technical races.
Why does elevation gain matter so much in training?
Elevation gain reflects the exact stressors of trail racing. Regular climbing and descending build muscular endurance, efficiency, confidence, and fatigue resistance.
Are lab tests like VO2max worth doing?
Lab tests can be useful for profiling athletes, but they should not drive training decisions in isolation. Field-based performance and training history are more relevant for ultra-trail performance.
What should coaches prioritise when training ultra runners?
Consistency over the years, terrain-specific exposure, elevation gain, and tracking performance trends matter more than chasing perfect physiological numbers.