[tech/medical] What is it with walking calories and MFP/trackers?

yirara
yirara Posts: 9,376 Member
Seriously, MFP grossly overstates calories for walking, and fitness trackers do the same. My Garmin estimated my calorie burn for a tiny 4km walk at 265 calories yesterday. My old Fitbit gave similarly crazy numbers. This can't be explained by gross vs net calories, neither with HR as my HR is substantially lower when walking compared to running and I've set zones to reflect this, based on HRmax estimate and a recent exercise test. Furthermore, there are apps out there advertizing these crazy burns as a way to lose weight quickly by just walking around the block every day. Btw, my running burns looks much more realistic. So what is it with walking? Is 0.3*weight*distance so wrong after all, or is walking just poorly implemented pretty much everywhere?
«1

Replies

  • sijomial
    sijomial Posts: 19,811 Member
    I use Strava on my phone (so no HR involved) just to record distance walked and the calorie estimates are ridiculously high. Correcting for net v. gross doesn't bring the estimates into what I would regard as reasonable so I use an online calculator.

    It shouldn't be that it's assuming I'm running either as I'm only walking at about 3.5 mph.

    Lazy programming perhaps using a running formula for both walking and running? Vanity calories?
  • yirara
    yirara Posts: 9,376 Member
    sijomial wrote: »
    I use Strava on my phone (so no HR involved) just to record distance walked and the calorie estimates are ridiculously high. Correcting for net v. gross doesn't bring the estimates into what I would regard as reasonable so I use an online calculator.

    It shouldn't be that it's assuming I'm running either as I'm only walking at about 3.5 mph.

    Lazy programming perhaps using a running formula for both walking and running? Vanity calories?

    I made the same observations with various apps just looking at distance and not HR, or even steps and stride length. Glad I'm not the only one here Sijomial. My running and cycling calories are on average not too bad if I substract BMI from what is already supposed to be net calories (I don't use Strava or a powermeter). Walking is ridiculous wherever I look. So... no idea.
  • RunsWithBees
    RunsWithBees Posts: 1,508 Member
    Lots of trial and error in the beginning when I used the mfp calculations, I found out pretty fast their counts are way off! I don’t know why the estimates are so off for other trackers either but I now only use the free iRunner app for walking, running and hiking and the calorie burn has been spot on for me. I ALWAYS eat back ALL the calories it claims I’ve burned from these exercises and over the span of 7+ years my weight has responded accordingly so I know it’s accurate for me.
  • fitnessguy266
    fitnessguy266 Posts: 150 Member
    OP, may I ask which Fitbit version you have? I've been using the Versa 2 for a couple of months now, and I've noticed from personal tests that it's pretty accurate in regards to the calories burned formula.
  • dragon_girl26
    dragon_girl26 Posts: 2,187 Member
    OP, may I ask which Fitbit version you have? I've been using the Versa 2 for a couple of months now, and I've noticed from personal tests that it's pretty accurate in regards to the calories burned formula.

    Same here..I also have the Versa 2 and eat back the majority of the calories it says I burn. My weight seems to be reacting accordingly as I would expect.
  • PAV8888
    PAV8888 Posts: 13,525 Member
    A "fitted" mean value will be accurate or near accurate for most, relatively not accurate (up or down) for some people, and extremely innacurrate for very few (but NOT zero) people.

    It sounds like you're one of the few.

    Comparing logged values and their expected results against the corresponding long term weight trend allows for a personalized correction.

    My own correction is of the order of 3.5% of TDEE or less when logging accurately and closer to 5.5% when logging my food intake more loosely.

    Even thought I exceed MFPs very active setting on the basis of deliberate walking I haven't found it needful to estimate how far off specifically walking is for me.

    I HAVE noticed that on relatively inactive days my result deviations from tdee are almost zero, so most of the tdee error I discuss above is probably due to an overestimate of activity calories if I were to go with my gut feeling. Then again maybe if I don't move I lose some water weight!

    It actually doesn't really impact me that much since on balance I would rather be active than inactive and the calories burned are tracked mostly for informational purposes and to ensure I have a cautionary flag waving in front of me when I hit the cookie jar!
  • Duck_Puddle
    Duck_Puddle Posts: 3,224 Member
    The .3 x weight x miles (and the .6 for running) came from a Runner’s World article that is no longer available. But it was based on a gross calorie burn of .5 x weight x miles - leaving the .3 as the net calorie burn.

    If I subtract my BMR from the .5 x weight x miles, I end up higher than .3. Some others put it to the test a few years ago and also came up higher than .3 (more around .4-.45 or so) as a net figure.

    That’s not going to make a massive difference in calorie calculations unless you’re walking a significant distance.

    But, the .3 doesn’t seem to really be correct across the board. Perhaps that’s where it lands for a certain weight range? Who knows.

    The .5 figure (at least for me) comes reasonably close to gross calorie burn based on METS.
  • spiriteagle99
    spiriteagle99 Posts: 3,673 Member
    I may be an outlier. For me the mfp numbers are about the same as those on my Garmin and I can eat 110% of the calories given and still maintain my weight.
  • nooshi713
    nooshi713 Posts: 4,877 Member
    I find my Fitbit walking calories to be very accurate. I don’t earn many calories for my steps though.
  • yirara
    yirara Posts: 9,376 Member
    edited March 2021
    I get more than twice for walking, thus in the range of 0.65 with any tracker. If I chose here very sliw pace walking it works, yet this is for my very fast walking pace -and very brisk again gives about double the calories.

    I don’t have proper walking data, but years of running data. I don’t wabt to believe that brisk walking, that I can keep up effortlessly until I run out of energy costs the same caliries as running. Which is hard for me
  • robertw486
    robertw486 Posts: 2,386 Member
    yirara wrote: »
    I get more than twice for walking, thus in the range of 0.65 with any tracker. If I chose here very sliw pace walking it works, yet this is for my very fast walking pace -and very brisk again gives about double the calories.

    I don’t have proper walking data, but years of running data. I don’t wabt to believe that brisk walking, that I can keep up effortlessly until I run out of energy costs the same caliries as running. Which is hard for me

    You may be walking quickly enough to be in the crossover area of efficiency. Walking very quickly can burn more calories than running slowly, but the exacts of the crossover points seem to be disputed even in various studies. But they all seem to agree that generally walking is most efficient at 3-4 MPH, and then efficiency drops and calorie burn goes up quickly with speed. Running efficiency is a much more linear progression, with efficiency increasing with speed.


    The Runners World article was based on a study that compared the two, and the paces used were right at about 3 MPH for walking and 6 MPH for running. IIRC one of the primary goals was to compare to various prediction models that were available. ACSM models proved to be one of the more accurate, but in some cases the prediction models were off over 30% IIRC. If any tracker uses the wrong prediction model, there may be more variance in accuracy based on body composition, speed, etc... essentially anything included in the prediction model.

    Also keep in mind that the tests showed that when adjusted for fat free mass, the gender differences more or less disappeared. But if a device uses a model based on averages for fat, being and outlier in the norm skews your data more. I recall that in the study the subjects for males tended to be on the leaner side for males, while the females were closer to an average fat mass. Somewhere around 12% for males and 24% for females.

    So plenty of places for data errors, even when you drill it down fairly deep.


    And it's possible that you are just an outlier to some extent. I worked with a woman that had a very brisk walking pace, to the point it was difficult to keep up with her without extra effort. She said she couldn't jog to save her life. It was just her natural walking pace as well. She was somewhat long legged for her height, but not super lean, and her stride was very long.
  • MarttaHP
    MarttaHP Posts: 68 Member
    I'm pretty sure my Garmin (FR245) also overestimates my walking calories. It's definitely much higher than using a MET-based estimate would give me.

    Bizarrely enough, if I walk on the treadmill instead, my watch underestimates the calories. I compared two 1-hour walks at the same pace (12 min/km) with the same average HR (77 bpm). The outdoors walk gave me 265 kcal, the treadmill one (recorded as a "Treadmill running" activity on my watch) 108 kcal. Obviously there was more variance in my HR outdoors due to hills and wind and whatnot, but that discrepancy was pretty huge.
  • yirara
    yirara Posts: 9,376 Member
    I walked 14km yesterday. There were a few stops for photos and playing a game, but anyway, Garmin gave me a total of 837kcal. Which is a lot for flat terrain. Though to be fair, my old Fitibit Charge 2 would have given me closer to 1400 :D I still suspect Garmin is actually giving gross calories even though the workout overview states net. But that's still a rather high overall.
  • Lietchi
    Lietchi Posts: 6,090 Member
    yirara wrote: »
    I walked 14km yesterday. There were a few stops for photos and playing a game, but anyway, Garmin gave me a total of 837kcal. Which is a lot for flat terrain. Though to be fair, my old Fitibit Charge 2 would have given me closer to 1400 :D I still suspect Garmin is actually giving gross calories even though the workout overview states net. But that's still a rather high overall.

    Where does it say that it's net calories? I'm sure the calorie burns given by Garmin are gross, judging from the chunk taken out of my calorie adjustment every time I take long walks, partially compensating the relatively high calorie burn of the walk itself.
  • rheddmobile
    rheddmobile Posts: 6,840 Member
    Mapmyrun gives me a huge count for walking....BUT I eat them back, have for three years, and am maintaining. So it seems to be accurate for me, at least.
  • sijomial
    sijomial Posts: 19,811 Member
    Mapmyrun gives me a huge count for walking....BUT I eat them back, have for three years, and am maintaining. So it seems to be accurate for me, at least.

    I would say that you eating your walking estimates are effective rather than accurate which would need particular verification of purely that activity rather than being just a part of a multitude of estimates that results in your calorie balance.
    But it's a good remender that accuracy is nice but it's not a requirement for successful weight control.

    I lost weight using what I later discovered were inaccurate gym machine, walking app and heartrate numbers for a large exercise volume purely because I was consistent and made adjustments based on results.
  • yirara
    yirara Posts: 9,376 Member
    Lietchi wrote: »
    yirara wrote: »
    I walked 14km yesterday. There were a few stops for photos and playing a game, but anyway, Garmin gave me a total of 837kcal. Which is a lot for flat terrain. Though to be fair, my old Fitibit Charge 2 would have given me closer to 1400 :D I still suspect Garmin is actually giving gross calories even though the workout overview states net. But that's still a rather high overall.

    Where does it say that it's net calories? I'm sure the calorie burns given by Garmin are gross, judging from the chunk taken out of my calorie adjustment every time I take long walks, partially compensating the relatively high calorie burn of the walk itself.

    Hmm.. I can't find it anymore now on the app. It might only show up once the activity gets transferred. Odd one as it now only says calories burned. But when I transferred exercise to MFP 2-3 weeks ago I got the whole chunk of calories in a workout and no correction.
  • PAV8888
    PAV8888 Posts: 13,525 Member
    edited March 2021
    Correctly working integration compares the TDEE values for the day that mfp and the alternative device believe to be true, and issues a single adjustment that continues to change value till midnight when it becomes final.

    Correctly working integration eliminates any double counting.

    Now my information is valid for Fitbit, which I use. I don't have a Garmin, though I believe they currently work the same with MFP (as compared to Apple watch which used to have issues).

    If you start messing around by entering specific exercises directly on MFP you are actually manually overriding the detection values of your tracker for the specific time period.

    To me this would mess with my attempt of having an independent evaluation of my burn absent my own input and feeling of how hard I FELT I worked, which could be more or less depending on my mood.

    Other potential sources of error are: GPS instability, stride length, terrain flatness or steepness, terrain consistency (bouncing track vs sticky mud), dominant vs non dominant hand.

    And.... outliers DO exist.... the little upper tail end of 8 billion, .. is still over 10 million

  • cwolfman13
    cwolfman13 Posts: 41,876 Member
    I think the important thing to remember is that it's all an estimate...the .3xWeightxMiles is an estimate...from my own data, I personally find it reasonable, but also quite conservative...which is better than being grossly over.

    I wear a Garmin Instinct and I'd say overall it's pretty reasonable where gross calories are concerned in total, and for individual, isolated efforts like going for a walk or a ride, etc. My Garmin typically give me in the neighborhood of 250 calories for a 3 mile walk. The .3 formula give me around 175...if I subtract my estimated BMR calories from 250 it actually comes out to 175.5...so pretty close on that front...any discrepancies I've noticed are when distance and time are increased or if my HRM on my watch starts doing weird things...this can happen in both directions. I went on a walk the other day and my watch told me that my average HR on the walk was 56 when in reality on the same walk, it's usually around 90-95 or so.