Calorie Burn vs Miles Walked

[benjamingrantdu]

I'm worried about the calorie burn estimate I'm getting from fitbit.  I am 6', 195 lbs.  Based on everything I've seen it seems like I should burn about 100-120 calories/mile. However, when I look at the calories burnt vs miles walked, I'm getting about 190 calories per mile walked. That seems really high! I worry that if I use this value I will gain wait instead of lose it. Am I Missing something?

[Heybales]

Is that miles walked what Fitbit says you did, or outside better estimate from say a GPS app?

Because based on stride length, especially if not corrected, their distance could be wrong.

Distance wrong, calorie burn wrong.

Taking a current workout where you know the actual distance and can figure out speed, this calculator is using formula that studies have shown to be within 4% of measured calorie burn. Select the Gross option, since that is what Fitbit would be reporting.

http://www.exrx.net/Calculators/WalkRunMETs.html

And this is exactly the type of formula Fitbit is using, except they may have the distance wrong.

You also could be walking in a manner that fools the impact sensor in the unit. Short but intense steps could make it think you are really moving faster, but there again, their distance would be off.

[benjamingrantdu]

I'm not using an alternative method to validate distance, I understand if distance is off caloric estimates would in turn be off.

I'm talking strictly about distance according to FitBit versus calories burned according to fitbit.  Both the values are raw data from fitbit, nothing else. If you look at the plot I gave fitbit is giving me 189 calories/mile with a baseline of about 1700 calories per day.  So it's saying I'd burn 1700 calories per day + 189 calories for each mile I walked. If I use the calculator you linked or any other similar tool, it's much closer to 100-110 calories/mile.  I'm 195 lbs. I walk at about 3 miles/hour. So if you put in 60 minutes at 3 mph for my weight on the calculator you provided then I'm supposed to burn 307 calories total, or about 102 calories/mile. According to my fitbit, for 1 hour of walking at 3 miles per hour I would burn 567 calories.  That's quite a discrepancy.  

Using only fitbit distance and calorie burn gives a direct indication of their mathetmatical conversion from miles to calories. They think I've walked 3 miles and they're saying I burned 567 calories. That means that fitbit is doing a much different calculation than the one provided by that calculator.  If fitbit was using that exact calculation than the fitbit calories and fitbit distances should align perfectly with that calculator. They differ substantially. 

[UVcat]

---

@benjamingrantdu wrote:

Using only fitbit distance and calorie burn gives a direct indication of their mathetmatical conversion from miles to calories. They think I've walked 3 miles and they're saying I burned 567 calories. That means that fitbit is doing a much different calculation than the one provided by that calculator.  If fitbit was using that exact calculation than the fitbit calories and fitbit distances should align perfectly with that calculator. They differ substantially. 

---

But I'm pretty sure the fitbit is not calculating calories from distance, it's calculating calories from steps. So, if your stride length is longer than fitbit's estimate, you could be walking more miles than fitbit thinks (and more than 3 mph) but it could still be correct about calories because it's using steps. Does that make sense?

Another effect (but I think it's too small to explain the full discrepency) is that fitbit is also putting your BMR into those calories. Are the calculators you're looking at doing that too, or are they calculating the calories burned above BMR for walking a particular distance? If it takes you about 20 minutes to walk a mile, then fitbit will be giving you your BMR calories for those 20 minutes + the extra calories from walking, right?

Hope this helps!

-c

[benjamingrantdu]

>But I'm pretty sure the fitbit is not calculating calories from distance, it's calculating calories from steps. So, if your stride length is longer than fitbit's estimate, you could be walking more miles than fitbit thinks (and more than 3 mph) but it could still be correct about calories because it's using steps. Does that make sense?

I think it's more likely fitbit is calculating distance from steps, and calculating mileage from that.  Either way, I would have to be walking nearly double what distance fitbit is reporting which maybe the distance is off by 10% but not a factor of two.

Regarding BMR, that should be accounted for in the baseline 1700 calories i burn per day.  I am getting an additional 200 calories/mile that I walk on top of my baseline of 1700.  So yes really if I walk for an hour fitbit's calorie calculation for that period of time is 1700/24 hours + walking calories, but since I'm looking at the slope of a line of calories versus steps, then BMR is the intercept (1700 which is very close to correct) and 200 is per additional mile.  Like the BMR for a given day is almost constant, so that get's subtracted out if I compare two days. One day I walk 4 miles, and the next I walk 6. The difference in calories burned between the two days is nearly 400 calories. Both days were 24 hours, so my basal calories burn is the same, the only difference is the additional calories due to walking.

Anyway, it's not a big deal, the calorie estimate just seems high to me. It would be comforting to me to know that if I kept my intake lower than my burnt calories that I would lose weight, but the fact that if I walk 4 miles one day and 6 miles the next day = 400 difference in calories, I'm skeptical that that is the case.  I'll just keep faith for now and see how it goes

[Heybales]

---

@UVcat wrote:

Another effect (but I think it's too small to explain the full discrepency) is that fitbit is also putting your BMR into those calories. Are the calculators you're looking at doing that too, or are they calculating the calories burned above BMR for walking a particular distance? If it takes you about 20 minutes to walk a mile, then fitbit will be giving you your BMR calories for those 20 minutes + the extra calories from walking, right?

---

For the calculator link above, Gross includes BMR, NET does not. In fact, it's the rare calculator that attempts to remove them.

But you are correct, that's what Fitbit, and actually a HRM, or database entry are doing, what did you burn in total for that block of time, including what would have been burned just resting. Well, estimated of course.

[Heybales]

---

@benjamingrantdu wrote:

I'm talking strictly about distance according to FitBit versus calories burned according to fitbit.  Both the values are raw data from fitbit, nothing else. If you look at the plot I gave fitbit is giving me 189 calories/mile with a baseline of about 1700 calories per day.  So it's saying I'd burn 1700 calories per day + 189 calories for each mile I walked. If I use the calculator you linked or any other similar tool, it's much closer to 100-110 calories/mile.  I'm 195 lbs. I walk at about 3 miles/hour. So if you put in 60 minutes at 3 mph for my weight on the calculator you provided then I'm supposed to burn 307 calories total, or about 102 calories/mile. According to my fitbit, for 1 hour of walking at 3 miles per hour I would burn 567 calories.  That's quite a discrepancy.  

Using only fitbit distance and calorie burn gives a direct indication of their mathetmatical conversion from miles to calories. They think I've walked 3 miles and they're saying I burned 567 calories. That means that fitbit is doing a much different calculation than the one provided by that calculator.  If fitbit was using that exact calculation than the fitbit calories and fitbit distances should align perfectly with that calculator. They differ substantially. 

---

Ok, I think I see the issue.

Are you going off a daily total, minus baseline resting, and figuring out calories / mile?

Like my baseline is 6.4 cal/ 5 min, or 1843 daily.

Yesterday Fitbit said I burned in total (with no manual corrected entries or exercise) 2984 calories.

So I got 1141 calories above resting for all my activity.

Steps was 12614 and Miles was 6.61.

So 173 cal/mile, above and beyond resting metabolism.

But that is not the same as walking for exercise, this is step and impact based.

Because actually that 6.61 miles was obtained with being awake 18 hrs, so that would mean 0.37 mph, obviously can't use that in calculator.

But using 3 mph, calorie burn per hour would be 183 NET, or 61 cal/mile. So same effect of almost triple.

What happens if you just look at the stats for a specific walking workout? Not what distance you racked up through the day.

Like for running, the worst estimate I've seen was about 20% short on calorie burn, 12% on distance, from what Fitbit gave for a specific workout.

But no, your example, as shown by my own figures being rather high too when compared to walking as activity, would have prevented me from having any weight loss, and it did not. Actually, Fitbit underestimates, because it gives BMR level burn to all non-moving time, when it really only applies to being asleep.

Being awake burns more, RMR.

Eating and digesting/processing food burns more, TEF.

Standing burns more but since no stes isn't seen.

So no, that's not the right way to view the math.

[benjamingrantdu]

By the way, thank you for your sincere and helpful responses. I was writing hurriedly and didn't show my appreciation before - so I apologize! Thanks for the input and again sorry for the rushed responses.

Honestly, I should be less uptight about it, I have lost weight so far so I shouldn't be complaining! 

[benjamingrantdu]

I'm not sure how you are concluding that is not the right way to view the math, could you elaborate. 

If I take five days that I sleep for nearly the identical amount of each day.  Day one I walk 1000 steps and burn 1800 calories. Day two I walk 3000 steps and burn 2000 calories. Day three I walk 5000 steps and burn 2200 calories.  Day 4, 7000 steps, 2400 calories, and day 5, 9000 steps, 2600 calories.  All of those have the same awake and sleep time and only differ in the amount of steps taken.

The RMR should be identical betwen those days. The only difference is the number of steps. For each 1000 additional steps I am getting 100 extra calories burnt, consistently, over long periods of time.  

[Heybales]

The RMR, rather, BMR used does not change, based on age, weight, height. So correct there.

But time of those steps is not in the picture by your math, you are trying to make it in the picture by assuming 3 mph.

These are not calorie counts based solely on distance, but impact of steps. But for straight exercise that has little variance and expected amounts of impact, I've found most see it match up, many times within 5%. But you start jump roping, or doing lunges with extra weight, ect, you'll get bigger calorie counts. Probably still underestimated for work done, but it does increase.

Assuming 3 mph the difference looks great, but why assume 3 mph? For example my 6.61 miles would only take 2.2 hrs.

So if all I did was sleep all day, and then for 2.2 hrs walk that 6.61 miles at 3mph, and nothing else.

1843 + 403 additional = 2246 calories. Using that walking calc with NET. Which Fitbit closely matches when I do a level walk.

But I did not just walk for 2.2 hrs.You add up all the time I have steps, it's more than 2.2 hrs.

4 sets of those steps was up and down 7 flights of stairs. Big impacts and calorie burn higher than flat steps, even though it doesn't know incline.

Many of the steps were mere shuffles.

Many were actually scooting around on my knees, talk about short steps.

So many of the steps were faster than 3 mph and worth bigger calorie count, some were slower. They are not all given the same calorie count, as evidenced by the 7 flights at lunch were faster and VAM time, before and after work were slower and AM time only.

Here are 5 spikes of steps and calories from my day for the 5 min increments shown in the log. This has removed the BMR calorie burn of 6.4 cal.

Calories / Steps = calories/step

31.1 / 443 = .0702 stairs up

39.0 / 549 = .0710 (VAM) down

33.9 / 477 = .0711 (VAM) up

30.9 / 429 = .0720 down

30.1 / 414 = .0727 to car

avg cal/step = 0.0714

Steps that day in total 12614

Calories for all the steps - 901

Base 1843 + 901 = 2744 (still shy by 240 from the 2984 I got)

Notice though the easier steps have more cal/step than more active ones. So start adding more those in to get a better average, I'm betting the math works out just right. I'm not going to add up all my above BMR calorie counts though.

Fitbit burn 2984 - 1843 BMR = activity burn 1141 / 12614 steps = .0905

Why would slower steps, or pace, burn more per step? Less efficient.

http://www.exrx.net/Aerobic/WalkCalExp.html

[benjamingrantdu]

Yeah I would totally understand that if what we were talking about was a 10-20% discrepancy.  

A couple things to consider. My walk patterns are not the same from day to day, but the graph I supplied is for 10 days and still has a very, very linear fit. The variations from the line are certainy due to exactly what you're talking about, different rates yield different caloric burn. No doubt about that. But as your link shows, that's the difference between, in a 100 lb person, 70 calories per mile and 60 calories per mile. I'm not doing any running so it's safe to just look at the walking graph.

Your example validates this.  Your calories/step vary from .07 to .073...less than a 10% difference in rate of calories. That aligns perfectly with the data I'm showing, my calories per step do vary but not by more than 10%.

However, the difference between what I'm seeing and what I would expect by assuming a constant calorie burn per step is 80%, not 10%.  Again, I get that my oversimplification could result in small fluxuation, I'm averaging nearly DOUBLE the projected calorie burn. That is like .07 cals/step to .14 cals/step...both your example and the graph you showed don't explain that at all.  And it wouldn't be like it sometimes crept up to .14 cals/step...it'd be like somehow i'm constantly at .14 cals/step.

Anyway, I agree, there could be some nuances that explain a little deviation. But the estimate seems off by a lot more than that to me. 

[Heybales · Accepted solution]

I should add, those 5 data points were my 5 highest steps/calorie counts.

Here are my 3 lowest, less 6.4 again. This is lower than the majority of my lower step counts, but these lower values are seen a bunch of times too.

2.5 / 24 = .104

2.3 / 18 = .128

2.0 / 11 = .182

So those majority of values do indeed increase the average up to where expected.

And again, if there was indeed this kind of gross over-estimation, then my minor calorie deficit should have been wiped out, and I should not have lost the last bit of weight.

I just don't think it's possible to take what are accurate formulas for walking as a specific activity and translating it to steps taken over the course of a day.

Now you start to get in to the proprietary formulas they have come up with to account for rest of the day, greater or lesser impacts seen, ect.

The research on 3-axis accelerometers is very interesting for when you have a known mass and assumed stride length, and the ability to judge free-fall time, acceleration, and therefore walking or jogging, some slight incline ability but not much, ect.

But that also starts with assumption the stride length is correct.

And I think there even the manual corrections may be done incorrectly, assuming the walking stride is purposeful exercise level walking, not the more easy pace most of the day is.

I'm betting the formula's for adjusting used stride length based on impact compared to setting of stride length, can only adjust so much. So if someone set their stride length to what they accomplish going 5 mph, then it won't adjust correctly on their daily walking of 1.5 mph say.

But that's why you look at results once past initial water weight loss, and see if weight loss matches up with difference between eaten and burned calories.

I personally used an adjusted height to compensate for higher LBM than average and higher RMR figure, and with that adjustment, my figures compared to actual loss during most times was 5%. The worst figures was when I had a lot of workouts that were probably off and manually logged.

[benjamingrantdu]

Those are great points. Thanks for the discussion.  I am happy to hear that it has performed well for you.  I'm shooting for about a 750 calorie defecit a day. So honestly if I think I'm getting a 750 calorie deficit and I'm off by 5-10% (which by your accounts sounds worse case scenario) then that wouldn't be a problem for me at all.  As long as I'm getting consistent loss that's all that I'm worried about. Previously I've been able to get a lot of calories burnt directly from exercising - e.g. run five miles - then I was able to mainly rely on calories burnt as a result of that + rmr. 

Anyway, I'm rambling. Point is my back is wrecked and I need to rely on fitbit + myfitnesspal to try and keep calorie defecit. I was up to 210 but at 195 now - need to lose at least 10 more.  Anyway, you made me feel much more secure about the process and I appreciate it.

Last question if you don't mind - do you combine with MyFitnessPal or do everything through fitbit? Thanks again!

[Heybales]

I would suggest if body is already under stress of injury, making sure the deficit is reasonable is even more important.

You might get a week or two out of 750 and really see 3 lbs loss, maybe 5 if you count in water weight, but I'd suggest backing that off to 250 for the last 5 lbs, if that is really 5 lbs down to reasonable healthy weight.

Otherwise that extra stress is just setting up a fight with your body you won't want, and likely won't win.

I do use MFP, but log exercise corrections on Fitbit.

Even though I've never had sync issues, nor even slowness between accounts (I don't use the apps at all really), logging exercise on Fitbit cuts back on 1 sync direction at least.

Since I just post my exercise on MFP anyway, I don't need the Exercise diary posting to show up, let it come across as adjustment from Fitbit.

I started on MFP so use the food diary much more with recipes and food items. I don't think I'd like Fitbit's at all.

But my deficits were becoming too much and impacting weight loss and performance, so wanted Fitbit for rest of the day.

On MFP under same username heybales.

[UVcat]

I agree that this is interesting. Got me curious, so plugged in my last ~2 weeks of calorie and distance stats (removing days where I logged a non-step-based activity) and got the following:

I'm 5'10", 140#. So, the calories/mile (109) for me is also ~50-60% above what one would expect for 50 calories/mile/100# (70 cal/mile). The BMR of 1600 calories also seems to be high by ~10-15%, not sure what that's about...

Yet, when I have logged calories I have found that, over appropriately-averaged time periods (months), I've lost (to within 10%) what I should have, using fitbit's burn numbers (e.g. I did a 250 deficit for 5 months and lost 10 pounds). So, I'm inclined to heybales's idea that the accelerometer's algorithms may be smarter than the number-crunching I can do at home on excel... makes me feel a bit better for spending the $$, honestly...

-c

[benjamingrantdu]

@UVcat great to hear that you experienced similar loss to what would be expected. 140 lbs is probably nothing I'll see ever again in my life, but I'll shoot for a passable BMI for now.  It's reassuring to hear that a defecit of 250 produced the intended results -- makes me feel more confident in the numbers.  I agree that I must be oversimplifying it and the people working for fitbit know what they're doing.

@Heybales thanks for the response. I will ramp down my deficit to 250 soon, you make a good point.  I use MFP the same was as you - food tracking there, fitbit for exercise. I used fitbit calorie tracking for about a week but was used to MFP and ended up going back. I log additional exercise in MFP but not for any good reason.  You make very good points about the difference in calories/step overtime and the fact that your results are consistent with fitbit predictions is very meaningful. Thanks for the great responses. 

[bmiller1969]

I know I am not burning over 400 calories when I sleep. The calories and steps taken are way off. Right now it is showing I have taken 1,130 steps and burned 570 calories. That would be awesome but I know it is not correct. I cant seem to get anyone to give me an answer for why this is??

[UVcat]

---

@bmiller1969 wrote:

I know I am not burning over 400 calories when I sleep. The calories and steps taken are way off. Right now it is showing I have taken 1,130 steps and burned 570 calories. That would be awesome but I know it is not correct. I cant seem to get anyone to give me an answer for why this is??

---

you burn calories all the time, just to stay alive (heart pumping, lungs breathing, all kinds of cellular processes), not just through purposeful activity. the burn number you see on your dashboard is a combination of your resting metabolism's calories burned (since midnight) plus whatever extra calories you've burned through activity. so, having 1000 steps at 8am is not going to result in the same number of calories burned for the day as having, say 1000 total steps at 8pm.

as for burning 400 calories while you sleep: my Resting Metabolic Rate (the amount of calories I would burn if i were sedentary all day) is about 1400. that means that if i sleep for, say, 8 hours, i burn over 400 calories during that time. if you are male, taller than me, heavier than me, or some combination of those, you will burn even more calories than that while you sleep. so, i wouldn't be so sure that you don't burn 400 calories while you sleep.

it is pretty easy to figure out if your fitbit is registering steps accurately: check your number of steps, walk for some amout of time while counting your steps in your head (say, up to 50 or 100) and then check your number of steps again. i did this while on a treadmill in order to calibrate my stride length and found that my flex got the same number of steps as i did counting in my head to about 1% accuracy. my counting is probably not accurate to better than 1% (especially when counting fast), so i consider my flex pretty accurate (under those conditions at least).

hope this helps,

c

[Mukluk4]

Double check your stride lengths for both walking and running.  If those are off, the caloric counts are incorrect also.