Your Watch Tracked 3,000 Calories — Why That Number Can't Build Your Fueling Plan

You finish your long run, glance at your wrist, and see it: 2,800 calories burned. Maybe 3,100. The number feels satisfying. It feels like data. And if you're like most endurance athletes, you've probably used that number to justify a post-run meal, estimate your calorie deficit, or even plan how many gels you'll carry on race day.

Here's the problem: that number is almost certainly wrong. And even if it were right, it still couldn't tell you when to take your next gel.

The Study That Confirmed What Runners Suspected

A study published in PLOS One in July 2026 put four of the most popular smartwatches through a head-to-head test against indirect calorimetry — the clinical gold standard for measuring energy expenditure. Fifty-two Apple Watch Series 8 users, 51 Garmin Forerunner 955 users, 50 Samsung Galaxy Watch 5 users, and 44 Fitbit Sense 2 users completed identical cycling protocols while researchers measured their actual calorie burn in a lab.

The results were not subtle.

Apple Watch came closest but still overestimated by an average of 22 calories per session. Samsung overshot by about 57 calories. Garmin was off by nearly 69 calories on average. And Fitbit? After the researchers removed seven wildly inaccurate readings, the average miss was still notable — include those outliers and it jumped to nearly 129 calories.

Every brand overestimated. None were within what you'd call "close enough" for nutrition planning.

The researchers also found that accuracy got worse as body fat percentage increased. If your body composition doesn't match the population the algorithm was trained on, the error compounds.

Their conclusion was direct: no current consumer smartwatch measures calories reliably enough for weight management or research purposes.

Why This Matters Less Than You Think — and More Than You Think

If you've been using your watch's calorie number to plan your daily meals or decide whether you "earned" a bigger dinner, the study says you should stop. That number is a rough estimate at best, and the direction of the error (overestimation) means you're probably eating back more than you actually burned.

But for race-day fueling, the calorie number was never the right metric to begin with. And that's the part most athletes get wrong.

Race fueling isn't about replacing the calories you burn. It's about delivering the right amount of carbohydrate to your working muscles at the right time, replacing the fluid and sodium you lose through sweat, and doing both at a rate your gut can absorb without revolt.

None of those things require a calorie count. They require different data entirely.

The Metrics That Actually Build a Fueling Plan

Your smartwatch isn't useless — far from it. It collects several metrics that are genuinely valuable for fueling decisions. The trick is knowing which ones to pay attention to and which to ignore.

Heart rate zones tell you how hard you're working. Higher intensity means faster glycogen depletion, which means you need carbohydrate sooner. A long run at 65% of max heart rate burns fuel differently than a tempo effort at 85%. Your watch tracks this accurately. The calorie estimate layered on top of it is where the error creeps in.

Training load tracks accumulated stress. Whether your watch calls it Training Load, Training Effect, or something else, this metric captures the volume and intensity of your recent sessions. Higher chronic training load increases your carbohydrate needs during exercise because a more trained athlete can oxidise more carbohydrate per hour. It also affects how depleted your glycogen stores are heading into a long session.

Pace and duration are the basics, and they're reliable. The longer you run and the faster you go, the more fuel you need. Your watch nails these numbers. Combine pace and duration with your body weight, and you can start to estimate carbohydrate needs per hour far more accurately than a calorie readout.

Sweat rate is the metric your watch doesn't give you — but should. Your watch tracks skin temperature, heart rate, ambient conditions, and workout intensity. From those, it can estimate sweat loss. Garmin already does this in some models. But sweat rate alone doesn't tell you what you lost in that sweat — sodium concentration varies enormously between athletes, from roughly 200 mg/L to over 1,500 mg/L. Two runners can lose the same volume of sweat and need vastly different electrolyte replacement.

The Gap Between Data Collection and Fueling Intelligence

This is the core problem. Your watch is an extraordinary data collector. It knows your heart rate second by second. It knows your pace, your elevation gain, your skin temperature, your training history. Garmin can even log your meals now, thanks to the nutrition tracking feature it launched in January 2026.

But data collection isn't intelligence. Logging what you ate and how many calories your watch thinks you burned doesn't answer the questions that actually matter on race day:

• How many grams of carbohydrate should I take per hour, given my body weight, pace, and training history?
• When should I take my first gel based on how fast I'm depleting glycogen?
• How much fluid do I need per hour based on my actual sweat rate in today's conditions?
• How much sodium am I losing, and is my electrolyte drink covering the gap?

Those answers require a model that sits between the raw data and the decision. Not a dashboard. Not a food diary. A system that takes what your watch already tracks and translates it into a specific, timed plan.

What to Do With Your Watch Data Right Now

You don't have to wait for your watch to get smarter. You can start using the data it already collects to make better fueling decisions today.

Stop using calorie burn as a fueling input. For daily meals, use it as a rough directional signal — nothing more. For race fueling, ignore it entirely and focus on carbohydrate targets per hour.

Pay attention to heart rate trends on long runs. If your heart rate drifts upward at a steady pace (cardiac drift), it often signals dehydration. Your watch shows this clearly. When you see drift start earlier than usual, your hydration plan may need adjusting.

Track conditions alongside your runs. Your watch logs temperature and sometimes humidity. Sweat rate increases roughly 10–15% for every 5°C rise in temperature above your comfort zone. If your summer long runs had higher average heart rates than your autumn ones at the same pace, some of that is heat-driven fluid loss.

Test your sweat rate manually. Weigh yourself before a run with no clothes on. Run for an hour at race pace with a measured amount of fluid. Weigh yourself again. The difference, plus what you drank, is your sweat loss. Repeat in different conditions to build a profile. It's not glamorous, but it's accurate — and it gives you a number your watch can't.

Use your watch's reliable metrics to build carbohydrate targets. A general starting point: aim for 60–90 grams of carbohydrate per hour for efforts over 90 minutes. Where you land in that range depends on your body weight, your pace (faster = more), and how trained your gut is. One standard gel contains about 20–25 grams. Work backward from your hourly target to figure out how many you need and when.

The Missing Layer

The September 2026 study confirmed what endurance athletes have suspected for years: the calorie number on your wrist is a rough estimate dressed up as precision. But the study also revealed something more interesting — even the best-performing watch (Apple Watch, off by an average of 22 calories) is measuring the wrong thing for fueling purposes.

Your watch already collects the inputs that matter. Heart rate, pace, duration, training load, environmental conditions. What it doesn't do — what no watch does natively — is run those inputs through a fueling model and hand you a plan.

That's what Sweatr does. It connects to your Apple Watch, Garmin, or Strava, reads the metrics your watch gets right, and translates them into a personalised hydration and fueling plan. Not a calorie target. Not a food log. A timed sequence of actions: drink this much at this time, take a gel here, adjust if conditions change.

Your watch is a sensor. It's a brilliant one. But a sensor without a decision engine is just a number on a screen.