Like many other first-gen devices, the Google Pixel Watch has a lot of quirks. So far, early users have reported that the watch dramatically overreported calorie burn due to a bug. According to Android Police, the Pixel Watch team is aware of the issue and suggests rebooting the device could fix the problem. That said, it’s an apt reminder that calorie burn isn’t a reliable metric.

In the Pixel Watch’s case, the bug appears to have impacted how a user’s basal metabolic rate was calculated. Basal metabolic rate, or BMR, is essentially the number of calories you burn every day simply existing. It’s calculated based on factors like your age, weight, sex, and height. If you put in the wrong data, you’re going to get the wrong calorie burn. According to Android PoliceFitbit’s software was starting off with incorrect user data, which was then fixed when the device was rebooted.

That’s annoying but it’s ultimately not that serious. That’s because you should never trust any wearable device to give you accurate calorie burn.

For starters, no two wearable makers use the same algorithm in determining how many calories you burn during exercise. They each use proprietary algorithms that consider factors like heart rate, accelerometer data, and your BMR. Different workouts will exercise different muscle groups, which will also impact calorie burn — hence why these devices have multiple sport profiles for tracking activity. For example, while running and cycling are both great for cardio, you burn slightly more while running because it uses more muscles.

There are also dozens of other factors that smartwatches don’t account for. Using myself as an example, I have polycystic ovary syndrome. People with this condition burn an average of 400 fewer calories a day than those who don’t. Nowhere on my smartwatch can I tick a box so the algorithm can account for that. My smartwatch also doesn’t know how much muscle mass I have, my fitness level, the medications I’m taking, or the thermic effect of the food I eat.

In short, a lot of people are never going to get an “accurate” number, no matter how diligently they log their exercise and food. At best, calorie burn via wearables can only paint a broad view of your long-term progress and activity levels — or a heads-up that something isn’t quite right. To drive the point home, a 2017 Stanford study found that out of seven trackers tested, none delivered calorie burn metrics “that were within an acceptable range in any setting.” The most accurate had an error rate of about 27 percent. The worst had an error rate of 93 percent.

This is the main reason why, as a wearables reviewer, I choose not to evaluate how “accurate” a device’s calorie burn is. I can’t. At least not in a meaningful way. What I can do — and what users should do — is check to see whether the metrics they get, including calorie burn, are consistent.

While testing the Pixel Watch, Iran three 30-minute runs on the same route, at around the same speed, using the Apple Watch Ultra and the Runkeeper app as controls. You can see the results in the following table.

They’re all in roughly the same ballpark. I have no idea which one is the right number, but it doesn’t matter so long as they’re all delivering consistent results. Which they are. If I wanted, based on these results, I could use the Pixel Watch to get a rough idea of ​​whether I’m in a calorie deficit or surplus, depending on my goals. Generally, so long as two devices are within about 1,000 steps and 500kcals for your total daily count, you’re not going to get a demonstrably different experience.

Now if the Pixel Watch had given me three wildly different calorie burn numbers for the same activities, that would be a problem. At that point, I’d dig a bit deeper. For example, when I reviewed the Samsung Galaxy Watch Active 2, it logged a walk that was off by roughly 11,000 steps and six miles compared to my control device. Other times, it delivered results that weren’t too far off. That kind of discrepancy is a red flag.

Of course, it’d be nice to have a gadget that’s 100 percent accurate. It’s also impossible, no matter what any tech executive says at a launch event. Wearables are meant to help you determine your baseline so you can visualize your progress over an extended period of time. Zeroing in on any one metric — especially one as indirect as calorie burn — is losing the forest for the trees.