Does Fitbit calorie burn get better over time?

[Heds28]

Hi, I’ve been training and dieting for a few year now and can control my weight very well, what I find is like most people is that Fitbit estimates the calories burnt far too high, this is fine I understand it’s a fairly simple equation used however I’m now using the Fitbit feature that logs your calorie in vs out and I’m wondering if it now sees my weight changes would it improve and get a better calculation on my calorie burn?

hope that makes sense?

basically it’s saying some days I burn 4000 calories, so when I eat 3000 I’m 1000 calories under and if that were true I’d see a huge weight loss when realistically I’ve only burnt around d 3000 and can maintain on that, so if it sees daily weigh ins vs food intake it should lower my calorie burn to correct this? 

[SunsetRunner]

I don't think there is any 'learning' involved in the algorithm.

It works fairly well for me, and seems to overestimate for my husband. I think the only thing you can do is to check that your information is correct (sex, weight, height, or whatever it takes into account). If you notice that the error is systematic (e.g. always overestimating by 1000) then I guess the best you can do is a mental correction.

[Dominique]

---

@Heds28 wrote:

I’m wondering if it now sees my weight changes would it improve and get a better calculation on my calorie burn? 

---

No, Fitbit makes no attempt to adjust its calories burned estimate based on changes in your actual weight. You are the one who needs to make the mental adjustment, now that you know by how much Fitbit overestimates (in your case) your energy expenditure.

[Heds28]

Bit of a shame really Fitbit has the data input there to do a rather accurate calculation, my data is all correct btw (would of been a nice easy fix)

[WavyDavey]

I don't understand the point. Losing weight doesn't have an immediate direct effect on calories burned. Are you saying because you weigh less, the calories burned should be reduced further? It's really about activity level more than weight as far as I know.

Edit: Make sure your weight information in your profile gets updated to your current weight. That will have an effect on the calories burned.

[Dominique]

@WavyDavey : what I think @Heds28 is saying is Fitbit should "learn" from his historical data (calories in, calories out and weight change) and automatically adjust calories out accordingly. For instance, if calories out (as estimated by Fitbit) are 4000, calories in (as logged per user) are 3000, but weight loss over four weeks is "only" 4 pounds (instead of the predicted 8 pounds), Fitbit should somehow thereafter report calories out as 3500 (assuming activity remains the same). I just don’t see how it would do it in practice. When should the adjustment kick in: after one week, two weeks, four weeks, on a daily basis?

[WavyDavey]

@Dominique, thanks for the explanation. Thanks makes sense. It would be pretty cool if the system could update that way.

[Colinm39]

I have found that Fitbit does adjust the calories burned value.. The difference is minuscule, unless I'm missing the point in this post.  Only 2 x sweet biscuits for 10KG.. But you have to keep amending your weight log.  I have set my Calorie Burn Goal to the weight I want to be.

I did some tests on my Fitbit One account, and found by dropping my settings by 10kg it reflected immediately on the outcome.  So I braved it, and I backdated my weight setting by 2 days on my Ionic because the effort factor was more meaningful with the HR recording.. The following is yesterdays effort.

I have used the Mifflin-St Jeor equations and the BMR is only 15 calories lower than calculating my sleep calories and equating for 1440 minutes.

Weight  BMR   Cal Burn Effort

89kg      1570     2559       1.62

79kg      1470     2397       1.63

========================

-10kg     -100    -162        +.01      Difference

[Heds28]

You guys have got what I’m trying to say. 

Basically if you eat 3000 and your weight doesn’t change the next day then you burnt 3000 calories that day, (I know different factors such as water/sodium etc can affect this) but it would be nice to try and get it a little more accurate, instead of around 1000 calorie out. 

I log my weight daily in Fitbit so it knows my cals in, cals out, weight, body fat %, age, height, sex, there’s online calculators that use this info and get it an awful lot more accurate. 

[JimC_99]

The only adjustment I have seen the Fitbit make based on weight is the decrease in the calculated BMR as you lose weight.  My Charge2 over estimates calorie burn by about 350 calories on average per day.  As others have commented, once you know how it tracks for your metabolism you can adjust.  Currently I am trying to maintain so I have set the tracker to a 250 calorie deficit and eat a bit under.

[Heds28]

Thanks for that, as an example I’ve maintained weight over the past few days but my in cos out difference has been 1337, 1193 & 1081 calories, rather large difference for me 

[gtg947h]

It's definitely frustrating that there's no way to manually input some sort of offset or correction factor.  I got around it by putting in a fake weight loss target of about 5 lb, and since my long-term daily error was in the 150-200cal range I put a daily deficit of -250.  It works out pretty well; I probably leave out a few things like condiments and often estimate quantities, especially at restaurants (and I've had a few non-tracked cheat days too) but I've stayed ±1lb for a couple months now.  I figure I can't be that far off.

I also wish the "calories remaining" feature accounted for the fact that I go to bed well before midnight every day, but I've learned to take about 200 off the number and it comes out at the end of the day.