3 ms·

I cannot comment on this case (and yet, here I am typing) but I will say this: I put up with the unreliable, fragile, jawbone up for years. I think I went thro

I cannot comment on this case (and yet, here I am typing) but I will say this: I put up with the unreliable, fragile, jawbone up for years. I think I went through about 5 or 6 models personally and got several for my family members. Why? Because it had, by far, the best sleep tracker of anyone on the market. I could finally see how much sleep I got and it had a huge impact on me. Fitbit has finally caught up about a year or so ago. (I was only tracking this as a consumer so not sure exactly). I would be very interested to know if the alleged theft was related to sleep tracking. Specifically, if it was around figuring out when the user is in REM sleep, deep sleep, light sleep or awake. Everyone else could not break out REM sleep and deep sleep.

dick_sucker2 8y ago
I'm really skeptical that it's accurate even now.
com2kid 8y ago
> Everyone else could not break out REM sleep and deep sleep. As someone who has worked with the engineers who implemented this on another product, anything except for the most broadly defined categories is exaggerating. We worked with multiple sleep labs to do controlled sleep studies on people. The way one of those sleep studies worked was a bunch of doctors looked over all the charts of data that had been gathered, and then voted on what type of sleep the subject was getting at different points in time! Give the same data set to a different set of doctors and you'll get a similar, but somewhat different, result from their analysis. With that in mind, it is not possible for a consumer grade wrist worn device to give absolute results. With just an accelerometer, you can get very broad categories, that are easily confused with someone sitting very still while watching the entire LOTR trilogy. Add in a HR sensor and you can get some slightly better data. Now not to say these results aren't good enough to be actionable. You'll notice differences in the graphs for making healthy lifestyle changes, heck I was able to see a before/after change on my data by skipping the after dinner gin and tonic, so existing consumer sensors can detect stuff. But anyone claiming whole lot of granularity is over stating their results. (Unless the state of the art has vastly moved forward in... oh 2 years or so, but given how messy all the sleep studies I saw were, I doubt it!)
- projectramo 8y ago
  Very interesting. I have noticed that my baseline REM (or should I call it “REM”) was different for the two (Fitbit and Jawbone) but if it’s lower than baseline on either device I feel dramatically worse than usual and if I get higher than usual REM I feel better. As you say, it’s detecting “stuff”

dick_sucker2 8y ago

I'm really skeptical that it's accurate even now.

com2kid 8y ago

> Everyone else could not break out REM sleep and deep sleep. As someone who has worked with the engineers who implemented this on another product, anything except for the most broadly defined categories is exaggerating. We worked with multiple sleep labs to do controlled sleep studies on people. The way one of those sleep studies worked was a bunch of doctors looked over all the charts of data that had been gathered, and then voted on what type of sleep the subject was getting at different points in time! Give the same data set to a different set of doctors and you'll get a similar, but somewhat different, result from their analysis. With that in mind, it is not possible for a consumer grade wrist worn device to give absolute results. With just an accelerometer, you can get very broad categories, that are easily confused with someone sitting very still while watching the entire LOTR trilogy. Add in a HR sensor and you can get some slightly better data. Now not to say these results aren't good enough to be actionable. You'll notice differences in the graphs for making healthy lifestyle changes, heck I was able to see a before/after change on my data by skipping the after dinner gin and tonic, so existing consumer sensors can detect stuff. But anyone claiming whole lot of granularity is over stating their results. (Unless the state of the art has vastly moved forward in... oh 2 years or so, but given how messy all the sleep studies I saw were, I doubt it!)

projectramo 8y ago
Very interesting. I have noticed that my baseline REM (or should I call it “REM”) was different for the two (Fitbit and Jawbone) but if it’s lower than baseline on either device I feel dramatically worse than usual and if I get higher than usual REM I feel better. As you say, it’s detecting “stuff”