Elevation Gain in Running: Grade Adjusted Pace Explained

You run your easy loop, the pace matches the plan — and somehow the session felt like a race. The answer is in the elevation profile: on a 5 % gradient the same pace costs up to 50 seconds per kilometre more than on the flat. Ignore elevation and you end up training too hard or too easy depending on the route. Grade Adjusted Pace (GAP) solves exactly this problem and makes hilly runs comparable.

What Grade Adjusted Pace actually calculates

GAP is a conversion factor: your real pace is translated into a flat-ground equivalent that reflects the same energy expenditure. It is based on measurements of oxygen consumption while running on a treadmill at different gradients — the best-known curve comes from Minetti and colleagues (2002) and underpins Strava, Garmin and most training platforms.

In practice: 5:00 min/km on the flat, 5 % gradient, actual pace 5:50 min/km — GAP will read 5:00 min/km. You ran exactly in the intended range even though the watch suggests otherwise. Without the correction you would assume you were 50 seconds too slow and put unnecessary pressure on yourself the next time out.

It matters to understand that GAP is a model, not a measurement. It describes the average runner, not you personally. Athletes who train off-road a lot are often better on climbs than the model predicts — and more cautious on steep descents.

How much time does one percent of incline cost?

The curve is not linear. In the moderate range between 0 and 6 % the extra cost is roughly 10 to 15 seconds per kilometre per percentage point. Beyond that every further point becomes disproportionately expensive, because the runner increasingly works against gravity rather than converting it into forward motion.

• 2 % gradient: about 20–30 s/km slower than flat — barely noticeable on a gently rolling loop.
• 5 % gradient: around 50–60 s/km. 5:00 becomes 5:50 — the classic case where an "easy" session feels like a tempo run.
• 8 % gradient: about 100–120 s/km. This is where running starts to lose out economically against brisk hiking.
• 12 % gradient: 150 s/km and beyond. At this point power-hiking is no longer a fallback but the better choice.

Downhill follows different logic: up to about 10 % descent the energy cost actually falls, because gravity does work for you — at 4 % downhill you gain back roughly 20 seconds per kilometre. Beyond that the curve flattens, and from about 20 % descent the eccentric braking work costs more energy than gravity provides. The rule that follows: you lose more going up than you gain going down. A profile with equal climbing and descending is always net slower than a flat route of the same length.

How to use GAP in training

The practical value is not in reviewing numbers afterwards but in dosing sessions correctly beforehand. Three applications matter:

• Control easy runs by GAP. In the base range GAP is a more precise criterion than heart rate, because heart rate lags on climbs and drifts with heat or fatigue anyway. A target of "5:30 to 5:50 GAP" means: uphill you may run 6:30, downhill 5:10. The effort stays constant.
• Convert intervals on hills. A threshold session of 4 × 8 minutes at 4:20 min/km on the flat corresponds to a real pace of 5:00 to 5:40 uphill depending on gradient. Forcing flat pace on a climb is not threshold work but VO2max work — in the wrong zone. Conversely, hill intervals are a clean tool for avoiding exactly that overload: the gradient slows you mechanically while intensity stays controlled.
• Account for load realistically. A 15 km loop with 300 metres of climbing is not the same as 15 flat kilometres. Training load (TSS) should include the elevation-corrected effort — otherwise you systematically underestimate every hilly session and wonder about unexplained fatigue.

Elevation gain in the training plan: realistic magnitudes

Elevation is not an end in itself but a training stimulus with its own dose. Useful guidance for weekly volume:

• 10–20 metres of climbing per running kilometre in normal training. At a 50 km week that is 500 to 1,000 metres — enough to develop specific strength endurance without ruining quality sessions.
• Flat marathon course: 200–400 metres per week is plenty. More would be a stimulus you never need on race day.
• Hilly marathon or half marathon: 600–1,000 metres per week, part of it at race pace on the climbs.
• Mountain marathon, trail, skyrace: 1,500–2,500 metres per week in the specific phase, including long climbs of 20 minutes or more.

The progression rule applies here too: no more than 20 % more than the previous week. And climbing does not belong in the recovery week — a deload week with 1,200 metres of ascent is not a deload week.

Uphill technique: short, upright, controlled

Running mechanics change measurably on a climb: stride length drops, cadence rises, the foot lands flatter and closer to the body, and the torso leans slightly forward. You should not fight that adaptation. Anyone trying to hold flat-ground stride length uphill either overstrides or loses cadence — both cost energy.

Three technique points pay off immediately: first, hold cadence rather than stride length. As the gradient rises the step shortens automatically — that is physiologically correct and energetically cheaper than a long stride. Second, bring the centre of mass over the foot instead of leaning back: an upright torso with a slight forward lean from the ankle keeps the hip angle open and allows full hip extension. Third, use your arms, but without tension — a strong arm drive stabilises the trunk and supports cadence.

And the descents? They are the underestimated part in exercise science. Downhill running is eccentric work: the muscle brakes while it lengthens. That produces high peak forces at low oxygen cost — and it is the strongest known stimulus for delayed-onset muscle soreness and for building muscle stiffness. So dose descents, do not maximise them. Two or three controlled descents per week during the build phase are enough. If your race includes 800 metres of descending, you should have experienced that load at similar length at least three times before — otherwise your quadriceps decide on race day, not your endurance.

Technique details in both directions are covered in Hill Training for Runners and Downhill Running Technique.

Where GAP reaches its limits

GAP is a good tool, but it has four known weak points you should be aware of:

• Technical terrain. On roots, rocks and slippery ground every step costs additional stabilisation work that the model does not include. A trail at 8 % is harder than an asphalt road at 8 % — GAP treats both the same.
• Very steep climbs. Above roughly 15 % the Minetti curve is no longer meaningful for most runners, because almost nobody runs there. If you are hiking, GAP is the wrong metric anyway.
• Altitude and heat. GAP corrects gradient only, not environmental conditions. At 25 °C or 2,000 metres of altitude the same GAP pace is considerably harder. You have to add those factors mentally.
• Smoothing of elevation data. Barometric altimeters are better than GPS elevation, but even they smooth away short ramps. A steep 200-metre ramp can appear as a gentle rise in the profile — and the perceived effort then no longer matches the number shown.

Practical check: three loops, one number

A simple test shows whether your training control matches reality. Pick three routes: one flat (under 5 m of climbing per km), one moderate (10–15 m/km), one hilly (over 20 m/km). Run all three at the same perceived effort, ideally within two weeks under similar conditions. Then compare:

• Raw pace: will be clearly slower across the three runs — that is normal and not a sign of fading fitness.
• GAP: should stay within 10–15 seconds per kilometre across all three sessions. If one session deviates strongly, the cause is almost always elsewhere — fatigue, heat, poor sleep, too little food.
• Heart rate: typically rises 5–10 beats in the hilly session even though GAP is identical. That is not a contradiction but the price of hill work — and a good reason to steer easy runs by GAP rather than heart rate.

Do this check twice per training block and you get a reliable feel for how your body responds to elevation — and you spot early when a session did not go to plan even though the numbers look right.

Common mistakes

• Forcing flat pace on hills. The classic error: the target says 5:00, so 5:00 is run on the climb too. The result is a threshold session where an easy one was planned — and three days of recovery needed.
• Treating elevation as a bonus. Adding 800 metres of climbing without adjusting volume raises load by 20–30 %, usually unnoticed because the kilometre count stays the same.
• Training only uphill. Running climbs is easy; the descent is the part that decides races. Never train descents and you pay twice on race day.
• Treating GAP as truth. The model describes the average. Your own uphill economy can deviate substantially — especially if you train mostly on the flat.
• Elevation in the recovery week. A deload week with hill runs is not a deload week. The stimulus is too high and the adaptation never happens.

How Peakora accounts for elevation

In the Peakora engine the elevation profile of your route feeds directly into the load calculation: a session with 400 metres of climbing produces a higher TSS than the same distance on the flat, and pace targets are corrected by the gradient factor — so you see from the start which pace uphill matches the intended stimulus. For hilly goal races the plan automatically shifts more volume onto the climbs and treats descents as their own stimulus. For mountain marathons and trails there is an additional specific phase with long climbs, in which weekly elevation rises on schedule while total kilometres drop slightly.

Frequently asked questions