Why your Whoop and your calorie app disagree, and which number to trust

Your Whoop and your calorie app disagree because they count different things. Whoop estimates your total daily burn, resting metabolism included, from heart rate. Most calorie apps start from a formula plus an activity guess, then add workouts on top. Both carry 20% or more error. The number to trust is your weight trend over 2-3 weeks.

That doesn’t make either tool useless. It means each one answers a slightly different question, and the trap is adding their answers together.

What is Whoop actually measuring?

Whoop estimates calories; it doesn’t measure them. Whoop’s support documentation describes a two-part model: a basal metabolic rate (BMR) from your profile (height, weight, age, sex), plus a proprietary function of heart rate once your heart rate rises.

Two things follow from that:

• The daily total includes BMR. A “3,200 kcal” day is not 3,200 kcal of activity. Roughly half of it may be the energy you’d burn lying still.
• Workout calories include BMR too. A one-hour ride logged at 800 kcal includes the 60-80 kcal you’d have burned on the sofa in that hour.

How accurate are wrist-worn calorie estimates?

Not very, in absolute terms. In a 2017 Stanford study (Shcherbina et al., Journal of Personalized Medicine), seven wrist devices were tested against indirect calorimetry while people sat, walked, ran and cycled. Most measured heart rate well. No device estimated energy expenditure with less than 20% error, and error was larger with higher BMI and darker skin tone.

A 2020 meta-analysis of 60 studies (O’Driscoll et al., British Journal of Sports Medicine) found accuracy varied widely by activity type. Combining heart rate with movement data reduced error. Whoop wasn’t in the Stanford study, but it uses the same class of heart-rate model, so expect the same kind of uncertainty.

Wearables are better at relative change than absolute truth. A day that reads 600 kcal higher than yesterday probably did cost more. Whether that’s exactly 600 is another matter.

What is my calorie app measuring?

Usually a prediction, not a measurement. Most calorie apps estimate resting metabolism with an equation such as Mifflin-St Jeor. Then they multiply it by an activity factor from the FAO/WHO/UNU physical activity level framework, typically 1.2 for sedentary up to 1.9 for very hard training.

The equation is well validated: a 2013 comparison found Mifflin-St Jeor the most accurate of the common equations. But it is still an estimate for an average person of your size, and the activity factor is a self-reported guess.

The other half of the app, what you ate, has its own error. In a classic New England Journal of Medicine study (Lichtman et al., 1992) of people who reported failing to lose weight on very low intakes, participants under-reported food by an average of 47%. That’s an extreme group, but missed oils, bites and portions are common in everyone’s logs.

Why do the two numbers disagree so much?

Mostly because of double counting. Here is a worked example for a 75 kg, 180 cm, 35-year-old male cyclist.

1. Resting metabolism (Mifflin-St Jeor): 10 x 75 + 6.25 x 180 - 5 x 35 + 5 = 1,705 kcal.
2. Calorie app baseline: he picks “moderately active” (1.55). 1,705 x 1.55 = 2,643 kcal. With a 500 kcal deficit, his goal is 2,143 kcal.
3. He syncs a one-hour ride from Whoop: 800 kcal. The app adds it, so his allowance becomes 2,943 kcal.
4. Whoop’s day total: 3,250 kcal.

The app’s 2,943 already counts the ride twice. The 1.55 multiplier assumes regular exercise, and then the ride was added on top. The 800 also includes about 70 kcal of BMR (1,705 / 24), which the baseline already covers. Meanwhile Whoop’s 3,250 could reasonably be 2,600 or 3,900 if its error is 20%.

Neither number is “right”. They are two estimates built on different assumptions, and stacking them compounds the error.

So which number should I trust?

Trust your weight trend, then use it to calibrate everything else. Your scale is the only tool that measures the result of what you ate minus what you burned, with no model in between.

The arithmetic, for the same rider over three weeks:

• Average logged intake: 2,900 kcal/day.
• Trend weight change: down 0.6 kg in 21 days.
• Energy gap: 0.6 x 7,700 = 4,620 kcal, or about 220 kcal/day.
• Actual maintenance: 2,900 + 220 = about 3,120 kcal/day.

Two caveats. The 7,700 kcal per kg figure is an approximation: Hall and colleagues show the real energy content of weight change varies, so use it over weeks, not days. And daily weight swings with water, salt and glycogen, so use a smoothed trend, not single weigh-ins.

How should I use my wearable, then?

Use it for the shape of your week, not the size of your plate. Wearables are good at telling you which days cost more: a hard ride versus a rest day, a big week versus a taper. Let the weight trend set the overall level, and let the wearable move the target up and down around it.

In practice:

• Don’t add wearable workout calories on top of an activity multiplier that already assumes training.
• Do let training days get more food than rest days, in proportion to what the wearable shows.
• Recalibrate every few weeks against the trend, especially after a change in training volume.

How Sweatmeal does this

Sweatmeal reads your burn from Whoop, Oura, Strava, Ultrahuman, Apple Health on iPhone (Health Connect on Android once it launches), and uses it only once: no activity multiplier stacked on top. It then corrects your maintenance estimate from your weight trend over time, so the target drifts toward what your body actually does. Each number is labelled measured or estimated. The full calculation is on our methodology page.