Basis smartwatch review It may have more sensors, but it doesn’t seem to do much with the extra data

I got interested in activity monitors about six months ago, in the early stages of a major push to reclaim fitness while I’m still middle-aged. (Mission accomplished, btw.) I wanted help counting calories burned. These devices try to calculate this for you by analyzing data from a motion-sensor strapped to your wrist, and the data can certainly be useful.

Their chief limitation is that they fail to detect exercise intensity that can’t be cleverly inferred from wrist jostling. Five minutes of squats will burn as many calories as twenty minutes of walking, but your activity monitor won’t have a clue, because — as far as your wrist-mounted motion detector can tell — you might as well be in a coma. But what if it knew your heart rate? What if it knew how much you sweated? Surely that would be better?

The Basis is reputedly the premium activity monitor on the market: the most sensors in the nicest package, a non-horrible looking watch (sure, it’s a little beefy, but big watches are currently fashionable). I upgraded to the Basis after about six weeks of bemused and exasperated experimentation with the simpler Nike Fuelband, which was baffled by all kinds of exercise that I do, particularly workouts at the gym, and soccer-like running while playing ultimate.1

So I bought the Basis. It has sensors on the back for heart rate and sweat. Cool.

Overall, the Basis does seem to be a much better product than the Nike Fuelband. Smarter, classier. More data, presenting more effectively. It has a (way) better iPhone app, better hardware, and better syncing. And yet I am still disappointed, surprise surprise: either I have worst luck with products, or I am much more demanding, or both.

Unfortunately, although it puts on a better face, the Basis is just as unreliable and flaky a service as the Fuelband: both erratically produce serious glitches in data. For instance, several times the Fuelband reported crazy numbers, clearly glitchy — more calories burned in a day than three people could go through if they were climbing mountains. And in the first three weeks, the Basis gave me at least three major discrepancies in calorie-burn reporting in different views of the data — again, just clearly glitchy. And that’s just the major, unambiguous problem. I have also encountered countless minor and temporary annoyances with both services. They are buggy and half-baked, period.

But what about the heart rate? And the sweat? Surely that data makes a meaningful difference? Can the Basis detect exertion with less motion?

Sweaty, heart-pounding action

The Basis fails at the main thing I wanted it for — everything that the extra sensors should be able to deliver, or what's the point in having extra sensors?

My Basis watch cannot detect the intensity of most activities other than walking, running and cycling ... or even the intensity of walking up a hill (which requires much more energy than walking on the flat). It detects increases in heart rate and perspiration, for sure, but seems unable to infer anything about calorie burn from that data (eg 2). This is so dramatically, obviously, and reproducibly wrong that I find it infuriating. What are Basis developers thinking? The data is right there — use it! If my heart rate and sweat spike while walking, I'm burning more calories!

I can only hope that Basis is using its richer data for other, subtler things.3

What about sleep?

Sleep tracking is presented as a major Basis feature. As high a value as I place on sleep hygiene — and I really do4 — I’m not sure what good this data is supposed to be. I was certainly intrigued by the idea of detecting sleep patterns I’m otherwise oblivious to, but it seems more like entertainment than health care.

And it’s a moot point, because the Basis doesn’t track sleep effectively. It failed an obvious test, disqualifying itself from my nights. Specifically, it failed to detect an obviously exceptional sleep.5 The Basis graphs of subjectively excellent versus poor sleeps were basically indistinguishable.

In fact, as far as I’ve been able to tell so far, the only thing the Basis gets right about sleep is roughly when it starts and stops — everything else fails to correlate with anything I can verify. It’s almost like it’s making it up.

The good

The bad

The ugly

  1. I typically jog a couple minutes to the gym, work out for 30 minutes, and jog home. The workout is easily triple the intensity of the jogging to and fro, but the graphed data from the Fuelband always showed basically … the opposite! Two prominent spikes of calorie burn on either side of a low plateau of exertion, graphically indistinguishable from the energy required to watch House of Cards. The Fuelband did a fine job with all the walks and runs, but it was getting tedious estimating how many calories I burned at the gym, climbing hills, playing ultimate, and other sweaty, heart-pounding, band-baffling workouts. ↩︎

  2. Walking up a steep hill, versus walking on flat. The watch certainly detects a much higher heart rate and perspiration for the former … but it reports exactly the same calorie burn, about 4-5 per minute, as the latter. ↩︎

  3. For instance, it probably uses it for more reliable autodetection of walking versus running versus cycling. Maybe. Because it does that really well. ↩︎

  4. I’m a champion insomniac, recovered mainly due to learning better sleep hygiene and sleep compression therapy, which I’ve written a lot about. If anyone really gets the importance of sleep consistency, it’s me. And yet I don’t really know what wearing a watch all night can do for me, no matter what it can detect. Awareness is good, and perhaps some people could use this technology to discover that they aren’t sleeping as well as they thought. But I’m already pretty vigilant about my sleep quality, by necessity. And there’s a risk that wearing the damn thing will annoy me and disturb sleep! And so it did, on some nights. Waking up irritated by a thing even once is a high price to pay. ↩︎

  5. I’d had a really great night — one of those super rare times when I barely move and have no sense of time passing. And yet my wife was up a lot — jet lag! — and reported that I was uncharacteristically oblivious to her movements. So here we have an almost ideal real world test of the sleep watch: an exceptional and objectively verified deep sleep, which should clearly contrast with data from previous nights. It didn’t. ↩︎