How Many Calories Does a Run Actually Burn?

Deriving the rule instead of remembering it

The American College of Sports Medicine publishes a metabolic equation for running: oxygen uptake in millilitres per kilogram per minute equals 0.2 times speed in metres per minute, plus 0.9 times speed times fractional grade, plus 3.5. The first coefficient is the oxygen cost of horizontal motion, the second the cost of vertical ascent, and the 3.5 is the resting term — the one metabolic equivalent everyone is carrying anyway.

The equation is written per minute, which is why the speed appears in it. Rewrite it per kilometre instead. At speed S metres per minute, covering a kilometre takes 1000 divided by S minutes, so the oxygen used per kilogram per kilometre is (0.2S + 0.9SG + 3.5) × 1000/S, which expands to 200 + 900G + 3500/S millilitres. The speed has cancelled out of the first two terms entirely and survives only in the resting term. Convert oxygen to energy at the conventional 5 kcal per litre and the gross cost is 1.0 + 4.5G + 17.5/S kilocalories per kilogram per kilometre. Drop the resting term and the net cost of running on the level is exactly 1.00 — no speed anywhere in it.

That is not a coincidence of the model. Margaria, Cerretelli, Aghemo and Sassi measured it on a treadmill in 1963 with two Olympic middle-distance runners, at speeds from 9 to 22 km/h and gradients from −20 to +15%. Their conclusion, in their own words, is that the net cost of running on the level is approximately 1 kcal per kilogram per kilometre and that the value is independent of speed and related only to the incline. They also noted that trained athletes were only 5 to 7% more economical than untrained subjects, and that running on the level costs about double walking at its most economical speed of 4 km/h.

Why running faster does not burn much more per mile

The table above runs the equation from a 12-minute mile to a 6-minute mile. Oxygen uptake nearly doubles, from 30.3 to 57.1 mL per kilogram per minute, which is exactly what your breathing tells you. But the cost per mile falls, from 0.825 to 0.778 kcal per pound of body mass — a change of about six per cent, in the direction most people do not expect. Running twice as fast burns roughly twice as much per minute and very slightly less per mile, because you are spending half as long carrying your resting metabolism around.

The practical consequence is that pace is the wrong lever for a calorie question, and distance is the right one. Six miles is six miles whether you take fifty minutes or seventy-two; the difference in the total is under fifty kilocalories, which is a mouthful of anything. If you want the number to go up, run further, run heavier or run uphill. Running harder makes you faster, fitter and more tired, but it does not meaningfully change the price of the route.

Gross, net, and the hour you would have lived anyway

The 3.5 in the equation is resting metabolism, and it is in the gross figure whether you run or not. One metabolic equivalent — 3.5 mL of oxygen per kilogram per minute — works out to 1.05 kcal per kilogram per hour, so a 155 lb runner is spending about 74 kcal an hour simply existing. An hour at a 10-minute mile covers 6 miles for a gross 753 kcal, of which 74 is resting — a little under a tenth. Go slower and the fraction grows: an hour at a 12-minute mile covers 5 miles for a gross 640 kcal, of which the same 74 is resting, nearly an eighth. If you are subtracting a run from a day's energy balance, the honest figure is the net one, because the resting share was already counted in your daily requirement.

There is a wrinkle in the resting term itself. The value equating one metabolic equivalent to 3.5 mL of oxygen per kilogram per minute was derived from a single 40-year-old man weighing 154 lb. When Byrne and colleagues measured resting metabolism by indirect calorimetry in 642 women and 127 men aged 18 to 74, the average came out at 2.6 ± 0.4 mL per kilogram per minute, or 0.84 ± 0.16 kcal per kilogram per hour. They concluded that the standard 1-MET value overestimates actual resting oxygen uptake by about 35% on average, and resting energy expenditure by 20%. So the resting correction is real but should not be applied too confidently in either direction.

Mass scales it, grade transforms it

Body mass enters the whole equation as a straight multiplier, which is why the rule is stated per unit of mass at all. On a 6-mile route the net cost is 526 kcal at 120 lb, 657 at 150, 788 at 180 and 920 at 210 — exactly linear, no curve. This also means the number falls as you lose weight, which is the quiet arithmetic behind a plateau: the same route gets cheaper as you get lighter, and a runner who has lost 20 lb is spending about 88 fewer kilocalories on every six miles than they were.

Gradient is the term that actually moves the answer. In the per-kilometre form the grade contributes 4.5 times the fractional grade to the cost. A 2% incline adds 9% to the net cost; 5% adds 22.5%; 10% adds 45%. Set that against the 6% the whole span of running speeds is worth and the ranking is unambiguous: a hill is a bigger variable than a pace, by a factor of several. One caveat. The 0.9 coefficient is validated for uphill running, and the equation should not be extrapolated downhill — descending costs less than the level, but not by the amount a negative grade would suggest, and the model has no term for the eccentric muscle work that makes a long descent so punishing the next day.

How exact is one kilocalorie per kilogram per kilometre?

Good to about ten per cent, and worth knowing which way it errs. The ACSM equation's net oxygen cost works out to exactly 0.200 mL of oxygen per kilogram per metre, which is where the round 1.00 comes from. When di Prampero and colleagues measured it directly in 36 amateur runners who had completed a marathon or half-marathon, the cost came out at 0.179 ± 0.017 mL of oxygen per kilogram per metre above resting, and — again — independent of treadmill speed. The ACSM figure therefore sits about 12% above the measured mean, roughly 1.2 standard deviations high. Converted, the measured value is about 0.90 kcal per kilogram per kilometre rather than 1.00.

There is a second, smaller wobble in the conversion itself. Five kilocalories per litre of oxygen is a convention that assumes a particular fuel mix. The true caloric equivalent runs from about 4.69 kcal per litre when you are burning pure fat to about 5.05 when you are burning pure carbohydrate — a spread of nearly 8%, and it moves during a run as glycogen depletes. Put the two uncertainties together and the honest statement is that running costs 0.9 to 1.1 kcal per kilogram per kilometre, which is exactly as much precision as the question deserves.

Watches, treadmills, and why exercise is a poor lever

Wrist devices are much better at heart rate than at energy. Shcherbina and colleagues put seven consumer wrist-worn devices through controlled laboratory activities — sitting, walking, running, cycling — with 60 participants. Median heart-rate error ranged from 2.0% for the best device to 6.8% for the worst. Median energy-expenditure error ranged from 27.4% for the best to 92.6% for the worst. The authors' conclusion was that most devices measure heart rate adequately but estimate energy expenditure poorly, and that caution is warranted in using those estimates. A watch reporting 640 kcal is telling you it is somewhere in the neighbourhood.

A treadmill display is a different quantity again. It reports gross energy, usually from a default body mass unless you have entered your own, and it is reporting the cost of running on a moving belt rather than over ground. Jones and Doust showed in 1996 that a 1% treadmill grade most accurately reproduces the energetic cost of outdoor running, because indoors there is no air resistance to push through. So a treadmill at 0% is a slightly cheaper run than the same speed on a road, and its calorie readout is a gross figure computed for a stranger.

Put the arithmetic together and one conclusion follows without any need for advice. A 155 lb runner covering 6 miles nets about 680 kcal, and Careau and colleagues, analysing total energy expenditure in 1,754 adults, found that energy compensation averages 28% — only about 72% of the extra energy spent on activity shows up as extra energy spent that day. So those six miles are worth roughly 490 kcal of added daily expenditure. Against that, a 12 fl oz regular lager is about 148 kcal, or 1.3 miles, and a gin and tonic about 160 kcal, or 1.4. Running is an extremely good way to become a better runner and a fairly weak way to move an energy balance, because the intake side moves faster than the legs do.

Frequently asked questions

Does sprinting burn more calories per mile than jogging?

Per mile, very slightly less. Per minute, a great deal more. The ACSM equation gives a gross cost of 0.825 kcal per pound per mile at a 12-minute mile and 0.778 at a 6-minute mile — about six per cent lower for twice the speed, because the resting share of the total is halved. Margaria's 1963 measurements found the same speed-independence up to 22 km/h. So if the goal is a number of calories, distance is the variable to change; pace changes fitness instead.

Why is my watch's number so much higher than this calculation?

Two reasons, and probably both. Your watch reports gross energy, which includes the resting metabolism you would have spent anyway — worth about a tenth of a slow hour. And wrist-based energy estimates are simply not very accurate: in a controlled study of seven devices and 60 participants, median energy-expenditure error ran from 27.4% for the best device to 92.6% for the worst, against 2.0 to 6.8% for heart rate on the same wrists. Use the watch for pace, distance and heart rate; use the arithmetic for calories.

How much does a hill really add?

A lot more than pace does. In the per-kilometre form of the ACSM equation the gradient term adds 4.5 times the fractional grade, so a 2% incline adds 9% to the net cost, 5% adds 22.5% and 10% adds 45%. Compare that with the roughly 6% that separates the slowest jog from a six-minute mile. The equation's uphill coefficient should not be extrapolated to descents: going down costs less than the flat, but not proportionally, and the model has no term at all for the muscle damage of a long descent.

Do I burn the calories from a run twice over, through afterburn?

No. Post-exercise oxygen consumption is real but small for the kind of running most people do, and it points the wrong way against a much larger effect running in the opposite direction. Careau and colleagues, analysing total energy expenditure in 1,754 adults, found energy compensation averaging 28% — only about 72% of the extra energy spent on activity turns into extra energy spent that day, because other components of expenditure fall in response. The net effect is that a run is worth somewhat less than its own arithmetic, not more.

Does the rule work for walking too?

No — walking is cheaper per unit of distance and, unlike running, its cost does depend strongly on speed. Margaria's 1963 paper states it directly: running on the level costs about double walking at its most economical speed of about 4 km/h, and the walking curve gets progressively steeper as speed rises, while the running line stays flat. That is why the running rule is so tidy and the walking equivalent is not: for walking you need the speed, for running you mostly do not.