Every gym in America has a sign above the treadmill that says some version of “running burns twice as many calories as walking.” The sign is correct. The sign is also irrelevant for most weight-loss outcomes, and this article is about why.
The per-minute math is not the question
A 165-pound adult walking at 3.5 mph burns roughly 4.0 METs of activity, which works out to about 280 calories per hour. The same adult running at 6 mph burns roughly 9.8 METs, which works out to about 685 calories per hour. So per minute of activity, running is 2.4 times more calorie-efficient than walking. That is what the treadmill sign is telling you, and it is true.
But the question a serious weight-loss researcher asks is not “how many calories does this activity burn per minute?” The question is “how many calories does this activity burn per year, given the actual probability that the user will still be doing the activity in twelve months?”
That is a different math problem, and the answer flips.
The dropout rate is the variable nobody puts on the sign
The published adherence literature on running as a self-prescribed weight-loss intervention is brutal. Between 50% and 70% of new runners stop running within the first six months. The injury rate is non-trivial — runner’s knee, plantar fasciitis, and Achilles tendinopathy together account for a sizeable share of dropouts that do not return to the activity. Self-prescribed running, especially for users new to high-impact exercise, has a quitter problem.
Walking does not have the same quitter problem. The twelve-month adherence rate for self-prescribed walking programs runs roughly three times higher than for running, in the populations where this has been studied. The injury rate is essentially negligible for moderate-pace walking on flat terrain. The activation cost of “I will walk for thirty minutes after dinner tonight” is dramatically lower than the activation cost of “I will run for thirty minutes after dinner tonight.”
The most efficient calorie burn per minute is irrelevant if you are not, in fact, doing the activity. Adherence is the variable that dominates the year-over-year math.
The annualized calorie math, illustrated
Take two hypothetical users, each starting a self-prescribed exercise program in January.
User A — Running. Five days a week, thirty minutes each session, at 6 mph. Per-session burn: roughly 340 calories. Per-week burn at full adherence: 1,700 calories. So far the math looks better than walking.
User B — Walking. Five days a week, forty-five minutes each session, at 3.5 mph. Per-session burn: roughly 210 calories. Per-week burn at full adherence: 1,050 calories. So far the math looks worse than running.
Now apply the adherence haircut. User A’s six-month dropout probability sits around 60%; by month seven, the per-week running burn is closer to zero. User B’s six-month dropout probability sits closer to 20%; by month seven, the per-week walking burn is still 800 calories or more (most walkers who continue past month six are still walking close to their original prescription).
Total burn over twelve months: User A clocks roughly 35,000 calories of running, almost all of it concentrated in the first five months. User B clocks roughly 47,000 calories of walking, spread across the full year. Walking wins the annualized calorie math by roughly 35%. The treadmill sign was right about minutes. The treadmill sign was wrong about years.
What the research actually recommends
The American College of Sports Medicine’s position stand on physical activity for weight loss recommends a minimum of 150 minutes per week of moderate-intensity activity for weight maintenance, and 250 minutes per week for weight loss. Walking comfortably meets both targets. Running can meet both targets — for the users who do not quit.
The 10,000-steps-per-day target, often dismissed as marketing arithmetic from a 1960s Japanese pedometer, has subsequently been supported by mortality data that puts the inflection point closer to 7,500 steps per day in adults under 60 and roughly 6,000 in adults over 60. The relevant point is not the specific threshold; the relevant point is that the dose-response curve for walking is essentially monotonic — more walking is better, with no meaningful injury penalty until very high volumes. Running’s dose-response curve is more complicated and includes a non-trivial injury cost above moderate volumes.
What we would tell a friend
If a friend who had been sedentary for years asked us what to start with as a weight-loss exercise prescription, we would recommend walking. The activation cost is the lowest, the injury risk is the lowest, the adherence rate is the highest, and the annualized calorie math works in walking’s favor. If, after six months of consistent walking, the friend wanted to layer in running — by all means. The order of operations matters: walking first, running on top of an established walking habit, never running as the standalone intervention for an untrained adult.
The most expensive calorie tracker in the world cannot fix the calorie-burn math of an exercise program that gets abandoned at week eleven. The single most reliable intervention we know is to pick an activity the person will still be doing in twelve months. For most adults, that activity is walking.