Another post about calories too high. Completely crazy.

[leifer20]

So many threads on this, so little attention by developers it seems. 

I have 5 years worth of data with a fitbit alta, NO HR. I can make a pretty simple linear model of steps to calories burned. In my case it is calories = 0.0676*steps+1992. This seems pretty reasonable given my BMR is 1700 or so. 

Two days with my fitbit inspire 3, I am now burning 0.111*(steps)+1732 calories. This is anywhere from 0-20% higher, depending on activity level. You can make an argument that if I sat in bed all day this model is accurate given my BMR. However, once I start moving (and getting my 10k steps in), this model goes in the trash. 

Why should active calories be counted? I haven't found a good reason. If you want people to have a good estimate of how many cals they are burning, USE THE ORIGINAL FORMULA WITHOUT HR!!! I have found it to be quite accurate in my 5 years with the fitbit Alta, which is now discontinued for fancier LESS ACCURATE trackers. Do I need to buy a fitbit Ace now?

[JohnnyRow]

You are free to go to Settings and turn heart rate off.  Or you could put in it Clip mode which also turns off heart rate.

You obviously believe calorie burned count is more accurate without heart rate tracking so feel free to do your thing without heart rate, and allow us to do ours with heart rate.

[leifer20]

Thanks for the alternative viewpoint. I have not yet found it to be super helpful. But maybe I am overlooking something? Is the utility of heart rate in the trends and continuous monitoring of HR rather than its contribution to calorie count?

[Dopovo]

@leifer20 

First of all congratulations to have identified an algorithm to match accurately your records, that's very neat.

As @JohnnyRow, mentioned the HR sensor based Fitbit devices do not use steps to estimate calories burned rather the HR (steps are only used as a motivational tool or once for gamification, now no more). This is universally considered a more scientific approach (steps are irrelevant to the effort put so you can agree that 10,000 steps made roaming around the house are very different than the same amount of steps done while running at a good pace) but of course the quality of the HR record play a huge role.

Now I think it is incorrect to say what is a best model as a comparison with a third sensor (at least) would be necessary. At this stage what we can say is that the 2 are different, not that one is good and the other bad.

I don't have experience on steps based calories estimation but my personal evaluation is that Fitbit overstimated the calories consumed and I base my statement in several ways:

- for myself and my daughter we experienced that if we eat the same calories (approx, say =/- 100 calories) than estimated as burned by Fitbit (Luxe in her case now, Ionic/Sense in my case) we tend to increase weight. In order to have a good balance I had to reduce the calories estimated by Fitbit for an amount of 600 calories for me and around 400 for my daughter.

- Comparison with several other trackers (Withings, Amazfit, Polar, Coros) where Fitbit provides a quite larger number than the others (in case of Sense the number is very high also during a run while Luxe looks more reliable hence it appears the problem is in the algorithm during daily tracking).

Sorry for long reading.

[SunsetRunner]

I totaly agree. In its options Fitbit is positioned as a weight loss tracker. Such large errors with calories are not acceptable.