Ultra Elevation Data: How to Read Vert
Ultra Elevation Data: How to Actually Read the Vert Numbers¶
Your watch records elevation gain on every run. Strava aggregates it. Your training app puts it on a chart. And almost nobody outside of pro coaches actually knows how to turn those numbers into training decisions or race-day predictions.
Here is the thing: for ultra and trail runners, elevation data matters more than mileage. A 20-mile run with 4,000 ft of gain is a completely different stimulus than a 20-mile run on a bike path. A race with 12,000 ft of gain is not just "longer" than a 50K with 3,000 ft — it is a different sport.
This guide walks through what your elevation numbers actually mean, the three ratios worth tracking, how to predict a hilly race finish without lying to yourself, and the limitations of the data your watch is feeding you.
What you will learn:
- Why "total elevation gain" alone is a useless number — and what to track instead
- The three vert ratios that actually predict ultra performance
- How to convert elevation gain into a realistic finish-time estimate using GAP
- Why your Strava and Garmin elevation totals disagree (and which to trust)
- How much vert per week your training actually needs
Why Total Elevation Gain Is a Misleading Number¶
Two runners post the same Saturday long run on Strava: 18 miles, 3,200 ft of gain. Same stimulus, right?
Not even close.
- Runner A ran 18 miles on rolling terrain — short climbs averaging 150 ft each, distributed evenly across the run. Average grade: about 3.4%.
- Runner B ran 18 miles with one 6-mile, 2,800 ft climb in the middle, then 12 miles of flat. Average grade on the climb: about 8.8%. Average grade overall: 3.4%.
The total number is identical. The training stimulus is not. Runner B trained their climbing muscles, mitochondrial density on steep grades, and the mental skill of grinding sustained vert. Runner A trained rolling-terrain rhythm and short-effort recovery. Both have value. They are not interchangeable.
This is why "I did 5,000 ft this week" tells you almost nothing on its own. The number you actually need to interpret elevation data is vert per mile — also called elevation density.
Vert per mile (ft/mi) — the only ratio that matters¶
Take total elevation gain and divide by total distance.
- 5,000 ft over 50 miles → 100 ft/mi (rolling)
- 5,000 ft over 25 miles → 200 ft/mi (hilly)
- 5,000 ft over 12 miles → 416 ft/mi (steep / mountainous)
Now you have a number you can compare to a race. Western States averages roughly 150 ft/mi. UTMB sits around 280 ft/mi. Hardrock is over 320 ft/mi. If your average training run is 60 ft/mi and you are toeing the line at UTMB, you have a specificity problem that no amount of weekly mileage can fix.
The Three Vert Ratios That Predict Ultra Performance¶
Ultra coaches and training scientists generally track three elevation-related ratios. Each tells you something different about race readiness.
1. Race vs. training vert per mile (elevation specificity)¶
This is the most important number for trail and ultra training. Take the race's vert per mile, then take your recent training's average vert per mile. Divide.
training ft/mi ÷ race ft/mi = specificity ratio
- < 0.5: Undertrained for race terrain. Your legs do not know how to climb or descend sustained grades.
- 0.5 – 0.8: Acceptable for first-time ultra finishers, but expect surprises.
- 0.8 – 1.2: Well-matched. You have rehearsed the stimulus.
- > 1.2: Over-prepared on grade. Useful if you live at altitude or train on stairs and race rolling terrain.
This is the number Uphill Athlete and Run Infinite both reference in their published ultra plans — most recommend hitting at least 50–80% of race elevation density during your peak block. (Run Infinite)
2. Weekly vert progression (the 10–15% rule, applied to vert)¶
The classic running progression rule says do not increase weekly mileage by more than 10% week-to-week. For trail and ultra runners, the same logic applies to vertical feet — and it is the variable that gets ignored.
You can hold mileage flat at 50 miles per week and still injure yourself by jumping from 4,000 ft of weekly gain to 8,000 ft. The eccentric load on descents alone — quads, patellar tendons, glute medius — adapts much slower than aerobic capacity. iRunFar's training volume coverage and most modern ultra coaches recommend thinking in time-on-feet plus vert, not just miles. (iRunFar)
A reasonable progression for someone training a mountain race:
| Week | Mileage | Vert |
|---|---|---|
| Build 1 | 45 mi | 5,000 ft |
| Build 2 | 50 mi | 5,500 ft |
| Build 3 | 50 mi | 6,500 ft |
| Recovery | 35 mi | 3,000 ft |
| Build 4 | 55 mi | 7,500 ft |
| Build 5 | 55 mi | 8,500 ft |
Notice: mileage barely moves in the last three build weeks. Vert is doing the work.
3. Peak week vert vs. race day vert¶
A common heuristic from Uphill Athlete and Trail Runner Magazine: in your peak training week (roughly 4–6 weeks out from a goal race), the total vertical gain should approach the race's total vertical gain, with about half to two-thirds of it concentrated in the weekend long run and back-to-back. (Uphill Athlete)
For a 50-miler with 10,000 ft of gain, that means your peak week is targeting ~10,000 ft of vert, with maybe 6,000 ft of it across Saturday and Sunday. If your peak week is 4,000 ft, you have not rehearsed the load.
See your elevation specificity ratio automatically computed from your last 4 weeks of Strava data on the NavRun analytics dashboard.
How to Predict Ultra Finish Time from Elevation Data¶
Road race predictors are simple: plug in a 10K time, get a marathon estimate. They assume one variable — fitness — because the courses are flat.
Trail and ultra races have two variables: fitness and elevation profile. Predicting finish time means accounting for both. There are three reasonable methods, in order of sophistication.
Method 1: Naismith's rule (and its corrections)¶
The oldest method, originally written in 1892 for Scottish hill walkers but still useful as a sanity check.
Add 1 hour for every 3 miles, plus 1 hour for every 2,000 ft of ascent.
For runners, the modern adapted version is: take your flat-pace estimate, then add roughly 30 seconds per mile for every 100 ft/mi of average grade.
It is rough. It ignores descent (which can be slower than flat running on technical terrain), altitude, and your individual climbing economy. But for a back-of-the-envelope check, it works.
Method 2: Grade-adjusted pace (GAP)¶
This is what Strava and most modern training tools use. The core idea: running uphill costs more oxygen per meter than running flat, and that cost can be modeled.
The classic formula, derived from D.B. Dill's research, says climbing 1 meter vertically costs about 1.31 mL of oxygen per kilogram of body weight — roughly the same metabolic cost as running an extra ~7 meters horizontally. (runbundle GAP Calculator)
To use GAP for race prediction:
- Look up the race's distance and total elevation gain.
- Estimate your average flat pace for the race distance (use a road race predictor).
- Adjust: add 15–20 seconds per mile for every 100 ft of average gain per mile.
So a runner with a flat 50K pace of 9:00/mi running a course with 200 ft/mi of gain should expect closer to 9:30–9:40/mi on average, before adding for technical terrain or altitude.
Method 3: Empirical ultra predictors¶
For 50K and below, GAP-based math is usually close enough. For 50-mile-plus races, especially mountain races, GAP under-predicts time because it does not account for:
- Cumulative quad damage from sustained descent
- Altitude (above ~6,000 ft, expect 2–4% slowdown per 1,000 ft)
- Technical terrain (rocks, roots, scree) — easily another 10–20% slowdown
- Aid station time and night-running pace decay
This is where ultra-specific predictors like RunSalty's Ultra Finish Time Predictor and Taktyk's pacing algorithm earn their keep — they bake in empirical data from major races (UTMB, Hardrock, Western States) rather than just GAP math. (RunSalty UFTP)
A practical rule of thumb most ultra coaches use: take your GAP-adjusted estimate, then add 10% for races over 50 miles, 15% for technical mountain races, 20% for high-altitude or 100-milers.
Race predictions in NavRun combine your fitness data, the race's elevation profile, and your recent training specificity to estimate a realistic finish. See it on the race predictions page.
Why Your Strava and Garmin Elevation Numbers Disagree¶
You finish a run. Garmin says 2,400 ft of gain. You sync to Strava. Strava says 2,150 ft. You upload the GPX to your training app. It says 2,580 ft.
Same run. Three numbers. None of them are wrong, exactly.
Where elevation data comes from¶
There are two ways a device can measure elevation:
- Barometric altimeter — measures air pressure and converts to altitude. Most mid-to-high-end GPS watches (Garmin Fenix, Coros Apex/Vertix, Suunto) include one. Sensitive, fast-responding, but drifts with weather changes.
- GPS-derived elevation — calculated from satellite triangulation. Available on every watch and phone, but inherently noisy with vertical error of 10–30 feet.
Strava's web interface, by default, discards your watch's recorded elevation and recalculates it from a digital elevation model (DEM) — basically a topographic map of the world. This gives more consistent numbers between users on the same trail but can disagree with your watch by 5–15%.
Which number should you trust?¶
- For day-to-day training tracking: use whichever your training platform uses consistently. Trends matter more than absolute numbers.
- For race comparisons: use the race director's published elevation gain if available. They usually report from the official GPX course.
- For long-term load tracking (weekly vert): be aware your numbers shift if you switch watches mid-cycle. A new Coros and an old Garmin will not produce the same totals on the same run.
The practical implication: do not chase a specific weekly vert target down to the foot. Track your trends, compare apples to apples (same device, same platform), and accept that elevation data has roughly ±10% noise built in.
How Much Vert Per Week Does Ultra Training Need?¶
There is no single answer, but here is a defensible framework based on race goals.
For a flat-to-rolling 50K (e.g., regional road or canal-path ultra)¶
- Target weekly vert: 2,000–4,000 ft
- Vert per mile: 50–100 ft/mi
- Focus: aerobic volume, durability. Vert is supplementary.
For a moderately hilly 50-miler (e.g., Bandera, JFK, many regional 50s)¶
- Target weekly vert: 4,000–7,000 ft in peak weeks
- Vert per mile: 100–200 ft/mi
- Focus: time-on-feet long runs with race-similar elevation density.
For a mountain 100K or 100-mile (e.g., Western States, Tahoe Rim, Bear)¶
- Target weekly vert: 8,000–15,000 ft in peak weeks
- Vert per mile: 150–250 ft/mi
- Focus: back-to-back long runs, descent durability, vert-specific blocks.
For a true mountain race (UTMB, Hardrock, Tor des Géants)¶
- Target weekly vert: 15,000–25,000+ ft in peak weeks
- Vert per mile: 250–400+ ft/mi
- Focus: power hiking, sustained climbing, multi-hour vert sessions.
These are rough brackets, not prescriptions. A 100-mile flatland runner moving up to UTMB needs a multi-year ramp, not a 16-week panic. (Trail Runner Magazine on chasing vert)
What If You Live Somewhere Flat?¶
The most common question from runners training for mountain races who live in Florida, the Midwest, or the UK lowlands. Real talk: you will be at a disadvantage versus runners with regular trail access, but you can close most of the gap with deliberate substitution.
Substitutes that work:
- Treadmill incline blocks — sustained 8–12% grade walks and hikes. Pre-race blocks of 60–90 minutes at 10–15% grade are standard practice for many Western States qualifiers in flat regions.
- Stair climbing — parking garages, stadium stairs, Stairmaster. Especially good for descent training, which treadmills cannot replicate.
- Weighted vest hikes — adds load that mimics climbing fatigue when grade is unavailable.
- Volume substitution — extra time-on-feet at lower intensity can partially compensate for missing the vert stimulus, though not the eccentric descent load.
What does not substitute: just running more flat miles. The metabolic and structural stimulus of climbing is different from flat running at the same heart rate.
Common Mistakes Runners Make With Elevation Data¶
After looking at thousands of training logs and race results, the same patterns repeat.
Mistake 1: Chasing weekly vert as a vanity metric. Vert is a means, not an end. Doing 15,000 ft per week for a 5,000 ft race wastes adaptation budget you could spend on speed, durability, or recovery.
Mistake 2: Ignoring descent training. Climbing makes you fit. Descending makes you finishable. Most ultra DNFs in mountain races come from quad failure in the second half, not lung capacity. (Run-Ultra: Are you doing too much elevation?)
Mistake 3: Counting "vert" from a stationary bike or stair machine the same way as running vert. Calorie equivalence is not training equivalence. Cross-training builds capacity without building running-specific elastic recoil. Useful, but not interchangeable.
Mistake 4: Reading a race's total gain without looking at the profile. A 50K with 5,000 ft spread across 30 rolling climbs is a different race than a 50K with one 5,000 ft climb. Always look at the elevation profile chart, not just the total.
Mistake 5: Trusting watch elevation data to the foot. Your weekly trend matters. The exact number does not.
Common Questions¶
Q: My watch says I gained 500 ft on a run that felt completely flat. Why?¶
GPS-derived elevation is noisy, especially for shorter runs. Small fluctuations get integrated into a positive total even on a true flat course. If your watch has a barometric altimeter and the weather changed during your run, that also affects the reading. Trust the trend, not the single data point.
Q: Is 50 ft/mi enough vert to count as "hill training"?¶
Below about 75 ft/mi, you are essentially running rolling terrain. Hill training in the structured sense usually means sustained climbs of at least 200 ft per effort, repeated, at controlled effort. Vert per mile is one measure; sustained climb duration is another.
Q: How do I know my race's true elevation profile?¶
The race director's published GPX file is usually the most reliable source. Strava's segment data can be misleading because of DEM smoothing on technical terrain. If multiple finishers from prior years report ~12,000 ft and the race says ~9,500 ft, the runners are usually closer to right.
Q: Does altitude affect how much I should train at elevation density?¶
Yes, indirectly. At altitude (above 6,000 ft), your perceived effort at any given grade increases because oxygen delivery is lower. You should target the same vert-per-mile in training, but be prepared for race pace to slow 2–4% per 1,000 ft of average race altitude above your training elevation. (Ultra Running Magazine: Training for Elevation Gain)
Q: Should I focus more on climbing or descending in training?¶
Both, but descending is more often neglected. Climbing volume builds aerobic capacity and muscular endurance. Descending volume builds quad eccentric strength and joint resilience — both critical in the second half of any mountain ultra. Aim for at least 30–40% of your weekly long-run vert to be descent-focused, even if it means running downhill repeats.
Q: How early before a race should I peak my vert?¶
Most ultra periodization models peak weekly vert 4–6 weeks before race day, with the highest vert week landing in that window. The final 3 weeks taper vert alongside mileage so you arrive fresh. Trying to bank a huge vert week in the final 2 weeks tends to backfire — you are training your post-race self, not your race-day self.
Q: My race has 8,000 ft of gain but I can only get to 3,000 ft per week. Am I undertrained?¶
Not necessarily. Specificity ratio is one input; total durability is another. If you have multiple long ultras under your belt and a deep aerobic base, you can compensate for a lower vert specificity by being patient on race day and pacing climbs conservatively. First-timers in this position should pick a less aggressive race or extend the training cycle.
Q: Is the elevation gain on Strava more or less accurate than my Garmin?¶
Strava's DEM-corrected elevation tends to be more consistent between users but can underreport on switchback-heavy or technical terrain. Garmin's barometric reading is more responsive but can drift with weather. Neither is "correct" — they measure differently. Pick one, stay with it, track trends.
Key Takeaways¶
- Total elevation gain is incomplete. Vert per mile (elevation density) is the number that lets you compare runs and races meaningfully.
- Three ratios predict ultra performance: training-vs-race vert per mile (specificity), week-over-week vert progression, and peak-week vert vs. race-day vert.
- GAP-based prediction works for under-50-mile races. Above that, add 10–20% for technical and altitude factors.
- Watch elevation data has ±10% noise. Track trends, not exact totals.
- Descent training is the most-skipped variable. It is what determines whether you finish a mountain ultra, not whether you start one strong.
Ultra running is a sport where the right variables matter more than more variables. Elevation data, used well, replaces guesswork with rehearsal — your race day stops being a surprise.
Start Reading Your Vert Numbers¶
NavRun pulls your full Strava history, computes your elevation specificity against your goal race, and shows weekly vert trends and per-mile density automatically — no spreadsheet required.
Connect your Strava account and your race in NavRun's analytics dashboard and your elevation data starts working for you, not against you.
Free forever for core features.