Using gross instead of net calories

Thanks for the awesome response! I've submitted a feature request and hopefully they'll make the change so I can compare my pre-Watch exercise to my post-Watch exercise.

Like many users, I'm also having issues with the accuracy of the "active calories". From what I can tell, the Workout app is using a seriously flawed model for computing activity. I've used a number of on-line calculators which agree with the Workout app being wrong. For example, I'm (almost) 54 and using a weight, time and heartbeat model for activity, I got a gross value of about 750 calories. The resting metabolic rate calories for that time period would have been 95 calories, in round numbers, for a net of 655. The Workout app reported it as about 370 net calories. My exercise calorie goal was 570 and let's just say I missed it very badly.

The other app I use (which I run in parallel) reported it as a 730 gross calorie workout. Again, netting the resting metabolic rate calories over the same time period would produce an active calorie value of 635.

Endomondo and the "weight + time + heartbeat" model are in very close agreement, while the Workout app value is just hopelessly wrong. I've used other exercise calorie models on-line and they pretty much agree with what Endomondo reports for gross activity. Since my resting metabolic rate doesn't change from app to app, they should all agree on net as well.

Thanks again for a most awesome response.

I did follow the "calibration" instructions and have done enough workouts now that it should have the correct values for the information it is using. The workouts I've used to conclude the calories are grossly inaccurate are ones that I've been doing for several years now. The reported information, such as average heart rate, distance, etc. are mostly correct. There is an error in the "flights of stairs" - from GPS data I know that the total change in elevation is equivalent to 7 flights of stairs, at 10' per flight. The distance is spot on the money with past workouts over the same course -- almost exactly 5 miles.

The scientific data show that active calories are based on non-fat body mass, regardless of sex, and my guess is that the calorie calculations make assumptions about non-fat body mass based on sex. This would produce incorrect results, though I don't know how much of an error that would produce. More accurate gross calorie values could use a metric such as weight and distance travelled over time, or weight and heartbeat over time. Either of those two gross calories values could then have the assumed resting metabolic rate calories subtracted from the gross calories burned to produce active net calories.

Thanks! I didn't know which sensor where was doing that calculation.

The iPhone does get the elevation change correct, and I'm assuming it uses GPS primarily for that. But when it converts the approximately +/- 70' change in elevation into "flights of stairs" it alternatingly decides that 70' is either one flight or two.

A quick followup after a few more weeks of using the Watch to track workouts.

Apple needs to do something to fix the "Active calories" calculation. There are multiple really good models for "Gross calories" based on weight and distance moved over time, as well as weight and heartbeat over time. Using those mathematical models I get values that agree much more closing to other applications, such as Endomondo or Polar.

As it stands now, none of the health tracking features of the Watch have any value to me. "Stand" can't seem to figure out that I'm standing and "Move" can't seem to get the calories right, and Apple doesn't seem to be responding.

@deggie -

It's pretty hard for the Watch to be correct when it says my heart rate is 50-something and I'm smack-dab in the middle of cardio **** and my actual heart rate is north of 130 or 140. Another watch owner I know has reported the same thing - the heart rate drops like a rock, as if there was some kind of small value unsigned rollover. My guess is that it's actually a sampling error caused by the sampling rate being well below what it should be. It smells like the sort of bug a software engineer would make that an electrical engineer would know better than making 😉

@deggie -

There are a lot of people reporting heart rate inaccuracies. It's not just myself and one other person. There's also only so tight the strap can be tightened before it doesn't get any tighter or it turns into a tourniquet.