Trust the Trend, Not the Calories: A Guide to Wearable Data

The 60-second version

• Heart rate can be useful for understanding exercise intensity and personal trends, especially during steady, rhythmic exercise.
• Steps can be very useful when we compare like with like: the same device, worn in the same way, across a sensible period.
• Exercise calories are estimates, not a food allowance and not a direct instruction to eat more or less.
• Garmin and Cronometer can show different totals because devices and platforms measure, estimate and combine information differently.
• Pillars makes decisions from the whole picture: your nutrition, rolling weight trend, movement, training, performance, hunger, recovery and real-world response.

What the numbers can—and cannot—tell us

Watch output	How Pillars uses it	Important boundary
Heart rate	A useful guide to exercise intensity and your own patterns when the signal and activity suit wrist measurement.	It can be less reliable during rapid changes, intervals, resistance training, high wrist movement or poor device fit. It is not a diagnosis.
Steps	A practical movement trend, normally viewed over 7, 14 and 28 days.	Different devices, wrists, wear patterns and algorithms can create a different baseline. Steps also miss some cycling, swimming, lifting and other non-step activity.
Active minutes and zones	Supporting context about how your week was organised.	Definitions vary between devices and should not be treated as identical across platforms.
Exercise calories	A low-confidence clue about session demand, considered alongside duration, intensity, heart rate where suitable, effort and recovery.	Not accurate enough to prescribe food from the displayed number.
Daily burn or total calories	Background context only, checked against what actually happens over time.	Not measured TDEE and not more authoritative than your longitudinal response.

Why Garmin, Cronometer and your watch may disagree

There are three different questions hiding behind one calorie total.

1. The device is estimating

A wristwatch uses sensors and an algorithm to estimate what happened. Accuracy can change with the person, the activity, device fit, body composition and the model or software being used. A watch can estimate heart rate reasonably well in one setting and estimate calories poorly in the same session.

2. The platforms may organise the data differently

Garmin, Cronometer and other systems may combine resting needs, baseline activity, imported exercise and tracker activity in different ways. A total can also become misleading if the same activity is counted twice.

3. Your body is not a static calculator

Even when exercise genuinely uses energy, the net change in your total daily expenditure may not match the exercise number exactly. Ordinary movement, efficiency, appetite and other parts of energy balance can change too. That is why two precise-looking totals can disagree—and why neither should automatically become your nutrition prescription.

Should I eat back my exercise calories?

Not automatically. Pillars does not use a 1:1 “eat back” rule. We also do not pretend that multiplying the watch total by 50%, or another universal percentage, fixes the problem. The direction and size of error are not consistent enough for that. There are times when extra fuelling is entirely appropriate. A longer run, demanding endurance block, high training volume or important performance session may require more energy and carbohydrate. We make that decision from:

• the session type, duration and intensity;
• your performance and training quality;
• hunger, energy and recovery;
• the wider nutrition plan;
• your rolling weight and measurement context where appropriate;
• what your body does across several weeks.

The training demand matters. The number on the watch does not make the decision alone.

How Pillars reads wearable information

We keep three things separate:

1. What was observed — the activity occurred, the device used, the period worn, a device change, or the fact a platform displayed a value.
2. What was estimated — calories, daily burn, active minutes and other algorithm-derived outputs. Step count is also an approximate device output, even when its trend is useful.
3. What we infer — the coaching interpretation after we combine the data with your nutrition, training, body-weight trend, performance, hunger, recovery and lived experience.

This stops an estimate quietly turning into a “fact”.

Trust the trend

A single day can be noisy. Rolling patterns are usually much more useful. Pillars will normally:

• compare the same device with itself;
• favour 7-, 14- and 28-day movement patterns over isolated days;
• look at completed, sufficiently representative days;
• keep training records visible alongside step counts;
• record a device, wrist, wear-pattern or major software change as a break in comparability;
• give your real response greater authority than a modelled total.

A watch can also be valuable simply because it encourages you to move. Prompts, goals and visible progress may help behaviour even when the calorie estimate is unsuitable for prescription. Behavioural usefulness and calorie accuracy are different questions.

If you change devices

Treat the first few weeks as a new baseline. Do not assume that 10,000 steps, 500 active calories or 45 active minutes on one device mean exactly the same thing on another. Tell your coach when you change:

• the watch or tracker;
• the wrist or placement;
• the normal wear pattern;
• the connected platform or synchronisation route.

We can then avoid joining two different measurement systems as though nothing changed.

When a reading looks wrong

Before reacting, check:

• Was the watch fitted securely?
• Was it worn for enough of the day?
• Was this an activity that wrist sensors capture well?
• Did the device or platform update?
• Has exercise been imported twice?
• Is the reading an isolated spike or part of a repeated pattern?

Unexpected heart-rate readings accompanied by symptoms, or a genuine health concern, need appropriate medical assessment. A consumer watch cannot diagnose, clear or rule out a health problem.

Your practical takeaway

Use your wearable to:

• notice movement and training patterns;
• support consistency and accountability;
• guide suitable exercise-intensity conversations;
• give your coach useful context.

Do not use it to:

• earn food;
• punish yourself for a low-burn day;
• make automatic calorie changes;
• compare yourself unfairly with somebody using a different device;
• treat a displayed total as measured truth.

Related Pillars guides

Explore our free guides and HRV guide.

Evidence behind this guide

This resource is a client-safe translation of the current Pillars evidence review. Useful public starting points include:

• Living systematic review of Apple Watch accuracy
• Umbrella review of consumer wearable accuracy
• Free-living step-count meta-analysis
• Current-device energy-expenditure study
• Free-living validation against doubly labelled water

Source guide reviewed: 9 August 2026.

Help putting this into practice

Want to put your tracking into context? We can look at your food, activity and everyday routines together through nutrition coaching in Oswestry or online across Shropshire. Start with a free conversation, or explore more free learning guides.