Are treadmill calorie readouts accurate?
Treadmill readouts are usually wrong, and usually wrong in the same direction for the same person. Not because manufacturers are dishonest — because the machine is doing a reasonable calculation with inputs it does not have.
It does not know what you weigh
This is the big one. The energy cost of walking or running scales almost linearly with the mass being moved, which makes bodyweight the dominant variable. If the treadmill never asked for yours, it assumed one — commonly around 70 kg (155 lb), sometimes 68 kg (150 lb).
| Your weight | What the readout does | |
|---|---|---|
| 55 kg | 120 lb | overstates by ~27% |
| 70 kg | 155 lb | about right |
| 85 kg | 185 lb | understates by ~18% |
| 100 kg | 220 lb | understates by ~30% |
So the machine flatters lighter users and short-changes heavier ones, which is precisely backwards from what either would want. And if you did enter your weight once, check whether it is still true: six weeks into a deficit the stored figure is stale, and the readout quietly drifts high as you get lighter.
Incline is handled inconsistently
Raising the belt adds genuine work — you are lifting your bodyweight vertically as well as carrying it forward — and at a 5% grade the cost of walking climbs steeply. The trouble is that machines disagree about how to account for it. Some implement the ACSM metabolic equations properly. Others apply a cruder multiplier, and a few barely adjust at all. Two identical sessions on two machines in the same gym can differ by a third.
The equations themselves also assume steady, continuous effort. Intervals, the first few minutes of any session, and sprint work all sit outside the conditions they were derived under.
The figure is gross, not net
Almost every machine reports gross energy: everything spent during the session, including the resting burn you would have had anyway. Over 40 minutes a 70 kg (155 lb) person burns roughly 45–50 kcal just being alive, so a 450 kcal readout contains about 50 kcal that were never extra. Add the full number on top of a daily total that already includes your resting rate and you have counted those twice — the same trap described in what 10,000 steps actually burns.
Handrails quietly break it
Holding the rails, especially on an incline, offloads part of your bodyweight and hands some of the stabilising work to the frame. Measured energy cost at a given speed and grade drops meaningfully. Belt speed and grade do not change, so the display carries on counting as though nothing happened.
If you need the rails to stay on at that incline, the honest move is to lower the incline.
Why the treadmill, your watch and the app all disagree
Because they are not three measurements of one thing. They are three estimates built from different inputs.
| Source | What it actually knows | How it tends to err |
|---|---|---|
| Treadmill | Speed, grade, time, an assumed or stale weight | High for light users, low for heavy ones; gross; blind to handrails |
| Wrist tracker | Heart rate, motion, your profile | Decent on steady running; drifts on intervals, hills and heat |
| Calculator or app | Your real weight, distance, duration | Sensible on average; knows nothing about this particular session |
Which to believe:
- Flat, steady running — the watch, as long as your profile is current. Heart rate tracks effort well in exactly this case, and its error, while real, is at least consistent.
- Inclines and hill programmes — the treadmill, as long as you entered your weight and kept your hands off the rails. It can see grade; your wrist cannot.
- For a deficit — none of them to the kilocalorie. Pick one source, use it consistently, and judge it against the scale over a month rather than against the other two.
What you must not do is average them, or log the session twice. If your watch recorded the run, typing the treadmill’s figure into a food app as well double-counts it. Caldown avoids this by reading active energy from Apple Health rather than a machine readout, so a session only ever exists once.
Getting closer
- Enter your weight, and update it as it changes.
- Subtract roughly 1 kcal per minute of session time to turn gross into net for an average adult.
- Keep your hands off the rails.
- Treat the display as a consistent relative signal. If it reads 20% high it probably reads 20% high every time, which still makes this week against last week a fair comparison.
For context on how little any single session contributes, see how many calories you should burn in a day — and you can put a figure on the number that does most of the work with the maintenance calculator.
General information, not medical advice.