Most runners and cyclists apply flat-course pacing logic to hilly races — and pay for it in the final kilometers. The instinct to maintain constant pace regardless of gradient forces disproportionate energy expenditure on uphills. Running at 5:00/km on a 6% grade requires approximately 35% more metabolic energy than the same pace on flat ground. An athlete who maintains constant pace through hills arrives at the summit in significant oxygen debt, with lactate levels 25–40% higher than if they had modulated pace by effort rather than speed.
Research from the University of Colorado (2019) demonstrated that effort-based pacing on a hilly half marathon course produced finish times 1.8–3.2% faster than constant-pace strategies — a difference of 1.5–3 minutes for a 90-minute half marathoner. The principle is straightforward: slow down slightly on uphills (where the energy cost per unit of pace is highest) and recover that time on downhills (where the energy cost per unit of pace is lowest). The challenge is quantifying "slightly."
The Grade-Adjusted Pace Model
Gradient affects running energy cost in a well-characterized, non-linear relationship. For every 1% increase in uphill gradient, running energy cost increases by approximately 3–4%. For downhill gradients, energy cost decreases by approximately 1.5–2% per 1% of decline, but only up to –6% gradient — beyond that, braking forces increase energy cost again. The practical application:
- 0–3% uphill: Allow pace to slow by 5–10 seconds per km while maintaining target effort (heart rate or RPE). The time loss is minimal and the energy savings are substantial
- 3–6% uphill: Accept 15–25 seconds per km slower than flat pace. Attempt to maintain flat pace on this gradient and you will accumulate lactate 30–40% faster than sustainable race effort
- 6–10% uphill: Switch to effort-only pacing — ignore pace entirely, maintain RPE 7–8 (race effort), and accept whatever pace results. Runners should expect 30–50 seconds per km slower than flat pace
- 0–3% downhill: Allow pace to naturally quicken by 5–10 seconds per km. Gravity does the work — let it
- 3–6% downhill: Pace will be 15–25 seconds per km faster than flat pace. Increase cadence by 5–10% and shorten stride to reduce impact forces
- Steeper than –6% downhill: Eccentric braking forces dominate. Maintain controlled pace, increase cadence to 185–195 spm, and avoid over-striding — the quad damage from aggressive downhill running in early race stages creates a disproportionate performance penalty in the final third
Power Meter Pacing: The Gold Standard for Hilly Courses
Running power meters (Stryd, COROS Pod) and cycling power meters eliminate the gradient-pacing guesswork entirely. On a hilly course, the optimal strategy is constant power — or more precisely, slightly negative power distribution (2–5% higher power in the second half). Power inherently accounts for gradient: running at 250 watts up a 5% grade and 250 watts on the subsequent flat produces grade-appropriate pace adjustments automatically.
For cyclists, the science is even more established. A 2017 study in Medicine & Science in Sports & Exercise found that variable power output on hilly courses — pushing 5–10% above flat target power on climbs and recovering 10–15% below target on descents — produced 1.5–2.5% faster overall times than constant power. This is because the speed-power relationship is non-linear: the same number of kilojoules produces more average speed when more of them are spent at lower speeds (uphills) where aerodynamic drag is minimal.
Heart Rate Lag and the Uphill Trap
Heart rate responds to effort changes with a 30–90 second delay, making it a poor real-time pacing tool on short, punchy hills. An athlete who uses heart rate to pace a 2-minute climb will overshoot target effort for the first 60–90 seconds (when HR reads lower than actual metabolic demand) and only see the true cardiac cost as they crest the summit — by which time the energy debt is already incurred.
The workaround: use heart rate as a ceiling rather than a target on hills. Set a heart rate alarm at 90–92% of maximum and treat it as a hard limit. If the alarm triggers on a climb, reduce effort immediately regardless of pace. This prevents the catastrophic overreach that occurs when competitive athletes try to "keep up" with faster climbers in the first half of a hilly race. The athletes who finish fastest on hilly courses are almost never the fastest climbers — they are the athletes who lose the least time on climbs while maintaining the most speed everywhere else.
Eccentric Damage Management on Descents
Downhill running involves eccentric muscle contractions — the quadriceps lengthen under load to brake against gravity. Each stride on a 10% downhill grade generates impact forces of 3.5–4× body weight, compared to 2.5–3× on flat ground. This eccentric loading causes microscopic muscle fiber damage that accumulates throughout the race and manifests as a dramatic pace decline in the final 20–30% of the course.
Mitigation strategies:
- Increase cadence: Higher cadence (185–195 spm on steep descents vs. 170–180 on flat) distributes braking forces across more, shorter ground contacts, reducing peak eccentric load per stride by 15–20%
- Lean forward: Maintain a slight forward lean from the ankles (not the waist) to align center of mass with the descent angle. This shifts braking from the quads to the glute-hamstring complex, which is more fatigue-resistant for eccentric loading
- Train eccentric tolerance: Include 1–2 downhill-specific sessions per month in race preparation — 6–8 × 90-second hill repeats focusing on the descent rather than the climb. Eccentric adaptation (the "repeated bout effect") reduces muscle damage markers by 50–70% after just 2–3 exposure sessions
Pre-Race Course Analysis: Know Your Elevation Profile
Study the course elevation profile and identify key features: total elevation gain, location and gradient of major climbs, descent profiles, and the gradient of the final 5K. Divide the course into flat, uphill, and downhill segments and set pace targets for each. The most important single data point is where the last significant climb ends — if the course finishes with a long descent or flat section after the final climb, you can afford to push slightly harder on that last climb knowing recovery pace follows. If the course finishes uphill, extreme discipline in the first half is essential.
The tool accounts for cumulative fatigue, so its pace recommendations for the final climb reflect the energy cost of all preceding hills — not just the gradient of the hill itself.