If you take part in organised runs, you may well have doubted your smartwatch accuracy. For example, I took part in a five-mile track run over the Easter weekend and my Garmin Fenix 5X reported 4.93 miles. So who’s cutting corners? You, or the smartwatch maker?
Thanks to the charming people at Which?, I thought for a moment that I finally had data to back up my hunch that Garmin is conservative while others stagger in the opposite direction:
Fitness brands | Marathon distance at 26.2 miles | Percentage difference (whether above or below 26.2 miles) |
Samsung | 22.55 | -14% |
Polar | 25.17 | -4% |
TomTom | 25.17 | -4% |
Fitbit | 25.43 | -3% |
Misfit | 25.69 | -2% |
Garmin | 25.69 | -2% |
Huawei | 25.96 | -1% |
Withings | 27.53 | +5% |
Apple | 28.05 | +7% |
Pebble | 30.68 | +17% |
But those figures are a little misleading. The discrepancies Which? discovered are based on the distance each watch predicts based on steps, rather than GPS.
“Which? test participants walk on a calibrated treadmill at 4.8km/h for 10 minutes and run at 9.0-10km/h for 10 minutes,” explains the organisation.
What’s interesting is that so many of the fitness brands clump in a bunch at the 25.2 to 25.7-mile mark. It’s enough to suggest that a) they all use similar algorithms to extrapolate distance from steps. And b) that their technology requires improvement.
Smartwatch accuracy: What about GPS ?
Even choosing a watch with GPS inside doesn’t guarantee results. Which? claims that some GPS-enabled watches can still be off by 20%.
You may be thinking, “But GPS should surely be 100% accurate?” Sadly, no. “GPS-enabled smartphones are typically accurate to within a 4.9m (16-feet) radius under open sky,” states the GPS.gov website. “However, their accuracy worsens near buildings, bridges, and trees.” It’s then up to algorithms to correct discrepancies, and clearly not all watches are equally good.
Which? is well known for its exhaustive tests and is kind enough to explain how it puts such watches through their paces.
“All fitness watches with built-in GPS are tested for accuracy on a 1km route,” it says. “The route includes dense trees and an underpass to add the challenge of potential loss of GPS signal. We even add a hill climb and compare the elevation data to Ordnance Survey topographical data to get an idea of the accuracy of the distance calculator.”
It goes on to explain that “fitness watches with advanced running capabilities”, which will include those from Fitbit, TomTom and Garmin, are also subjected to a 5km run test.
Smartwatch accuracy: Which to buy?
So what does it mean if you’re thinking of buying a watch? Certainly, these results cast doubt on the accuracy of both Apple and Samsung if you switch off GPS to save battery life, and that’s something you may well decide to do on a long run like the London Marathon. It would put me off both brands: it’s no good thinking I’m running at an eight-minute pace when actually I’m doing 7:40 or 8:20 miles.
But, aside from its occasional underreporting, I’m very happy with my Garmin 5X. It’s a brand I’d be happy to recommend to anyone looking for a running watch.
READ THIS NEXT: Garmin 5X review: How good is it for runners?
Story updated 20 April to clarify the results were based on steps, not GPS.
In my experience and for my walking traits, the numbers don’t seem accurate for Misfit. I always get between -15 and -30% accuracy when I’m walking. Good for me as im compensating an undercount but not too good for a step counter.