I didn’t think I was going to do it, but I did. I bought a Fitbit. My Zeo died right about the time they went out of business; presumably as soon as they stopped supporting their ap it just disintegrated. Which is a shame, because although buggy, it remains the best sleep tracker I’ve used.
The Fitbit, by comparison, is pretty poor. Looking at my graph from last night, I appear to have fallen asleep at 1.05am, fifteen minutes after I went to bed. I know this is not the case, as I made it through an entire episode of QI before falling asleep. I also started using the Sleep as Android ap again a few days ago, to compare the two. It thinks I fell asleep at 2.30am, which sounds more accurate to me, as I’m pretty sure I remember seeing some time after 2 the last time I looked at the clock.
I actually changed the sensitivity setting on the Fitbit last night before bed to see if that would sort things out, but while the time I supposedly fell asleep remained way out, I also spent half the night “restless” and it’s only logged four and a half hours of sleep, which is much too far to the opposite extreme. I don’t remember waking up during the night at all, and although I know I move around a fair bit, I honestly can’t imagine I spent literally half the night tossing and turning. Sleep as Android certainly doesn’t think I did.
This makes me suspect that Fitbit is using a very crude algorithm for sleep, which only measures time in quite large blocks. So when I was lying still watching TV for say, ten or fifteen minutes, it thought I’d fallen asleep. And apparently, when I’ve tossed and turned a little, it’s logged me as restless for multiple ten or twenty minute periods. Which is nuts.
There’s a similar problem with it as a pedometre. I wear it on my left hand, but I’ve had to tell it I wear it on my right (dominant), because it thought I’d walked 300 steps when in fact I’d been having lunch and, presumably, gesticulating. I use my hands a bit when I talk, but I don’t think I’m too over the top. Even since I’ve changed it, I have a feeling it’s overcounting. With my old pedometre 5,000-6,000 steps a day was about my average, and I had to try really hard to get over 10,000. My average on the Fitbit is 8,000 or 9,000, and I’m reasonably certain I’m not that much more active, although I suppose as the weather’s clearing up I might be outside a little more.
The steps don’t really matter; as a friend pointed out, I can just work out how many it thinks I’ve done on what I consider a pretty active day, and aim for that. The sleep stuff is annoying, though, and as I already had a (dumb) pedometre that I was happy enough with, on the whole I’m not that thrilled with my purchase. It’s cool to have a record of my steps without having to enter them into a spreadsheet each day, and the fact that I usually wear it to bed means I forget it less often. I might start using the food diary to keep better track of coffee etc, but I feel like I’m trying to find a silver lining here. The fact that I can’t even export my own data into a csv without paying is ridiculous as well – given that Zeo offered that as standard, foolishly, it never even occurred to me to check whether Fitbit did as well. All in all, pretty disappointing. I couldn’t recommend it.