Well Fitbit HR ... You are wrong!

Jeniccm
Jeniccm Posts: 44 Member
edited November 30 in Social Groups
I love my Fitbit HR and it has been a huge motivator to lose weight and still will continue to be.

I started on this journey to lose weight on January 1st. My Fitbit is linked to MFP. I have trusted the calories burned quite faithfully but today I decided to do some math

I made an excel spreadsheet to track my Fitbit calories burned vs. Calories eaten. My calories are definitely accurate and I measure to the gram.

I added up my total deficit between the 2 figures and divided by 3500 to come up with a figure of 5.25lbs that I should have lost from January 1 to today. I've actually lost 10lbs. Which means Fitbit is underestimating my calories burned by quite a bit.

I've found some online calculators and got figures on my personal heart rate zones. I'm curious if the difference is that my resting heart rate is around 56 and Fitbit is calculating it's heart rate zones based on an average for my age which would be around 70? I customized this and will see if it starts giving me more of a burn.

All in all I feel like this makes me see that I can EAT MORE :dance:

Replies

  • T1DCarnivoreRunner
    T1DCarnivoreRunner Posts: 11,502 Member
    Assuming no water weight differences, that is roughly 200 calories per day under-estimated by FitBit. Did you see my comment in the other thread in this group about the recently published study? The study found that fitness trackers underestimate calorie expenditure and your results fit within the same range found in the study, though the study used more precise methods to calculate actual calorie expenditure.
  • BarbieAS
    BarbieAS Posts: 1,414 Member
    Yes, wrong, but not exactly WRONG.

    As noted in the comment before mine, there's going to be a range of accuracy when you use a device like a Fitbit. Not because the Fitbit isn't designed or functioning correctly, but because the calculations for calories burned are based on averages and estimates. Two people who are the exact same height, sex, weight, and age might have BMR/RMRs that differ by up to several hundred calories per day - which translates to possibly an even bigger difference based on activity.

    (It seems like you aren't upset about the difference you noted, so you probably already realize this, just noting it for others who might read this thread.)

    Congrats on being on the "good" side of that range! :) Though, don't forget about any water weight you might have lost at the beginning. I do these sorts of Excel sheets all the time, and if you're looking to get a true *actual* calories burned, you're better off excluding the first 1 or 2 weeks - you get a much more accurate picture that way. You might find that your TDEE isn't off by as much as you thought.

    Re: custom heart rate zones - My understanding is that Fitbit still uses its own heart rate zones (rather than your custom zone) to determine how many calories it believes you're burning during a period of activity. I don't know how resting heart rate plays into that from an actual calories burned/biological point of view, but I do know that changing your zones to a custom zone isn't going to change your calorie burn.
  • heybales
    heybales Posts: 18,842 Member
    Dittos - easily 2-3 lbs, if not more the first couple weeks was water weight lost. That doesn't count in math.

    From there it's actually pretty decent.

    What if you run the figures from 2nd week in until recent weigh-in.

    Really good results though.

    I had it nail with 3% for a 6 week period of training where my records were very accurate. 3% better than allowed inaccuracy on food nutrition labels, so I figure that's great.

    And no - HR zones has no bearing on calorie burn during exercise. That's merely for your training or review.

    About 2 weeks after starting use your lower restingHR should have been used to get better workout calorie burns though as it adapted better to you.

    But if very active during the day with walking around - your stride length could have a bigger bearing on the inaccuracy.
  • BarbieAS
    BarbieAS Posts: 1,414 Member
    edited March 2016
    heybales wrote: »
    Dittos - easily 2-3 lbs, if not more the first couple weeks was water weight lost. That doesn't count in math.

    From there it's actually pretty decent.

    What if you run the figures from 2nd week in until recent weigh-in.

    Really good results though.

    I had it nail with 3% for a 6 week period of training where my records were very accurate. 3% better than allowed inaccuracy on food nutrition labels, so I figure that's great.

    And no - HR zones has no bearing on calorie burn during exercise. That's merely for your training or review.

    About 2 weeks after starting use your lower restingHR should have been used to get better workout calorie burns though as it adapted better to you.

    But if very active during the day with walking around - your stride length could have a bigger bearing on the inaccuracy.

    I'm not totally sure that that's true, just based on my own research and experience. (I'll try to make a very long story short.) Based on extensive "actual" TDEE calculations like the ones we're talking about and then confirmed with an RMR test performed by an MD, my BMR is significantly lower than what Fitbit uses to calculate total calories burned. In order to account for that, I changed my age on Fitbit to something that better approximates what I'm actually burning. (I tried it with height, but decided the loss of accuracy on my stride and my body fat from my Aria scale was too much.) In order to get it correct, I had to change my age from 34 to 116 (plus still take 2 inches off my height). When I had my Charge HR, my workout zones started at like 60 BPM due to the age I had entered. I have a low resting heart rate (upper 50s), but not that low. So, my Fitbit often noted that I was in exercise zones even when I was just strolling through the parking lot and stuff, and it still seemed that my TDEEs were inflated. I changed to an Alta about 2 weeks ago when it came out - same activity, same settings, worn in the same way, only change is the loss of HR tracking - and my calorie burns are literally HUNDREDS of calories lower than they used to be - seems much more accurate. So, I can only conclude that heart rate was definitely taken into consideration when Fitbit calculates calories. (I did set a custom heart rate zone that was more accurate for my real age, and it didn't change anything.)

    I can't find anything in the official Fitbit info that confirms this, but it does seem to have at least been confirmed by the Fitbit community moderators. https://community.fitbit.com/t5/Charge-HR/Calories-Burned-And-Heart-Rate-Monitor/m-p/1078376/highlight/true#M75243
  • heybales
    heybales Posts: 18,842 Member
    Well, age is used in default zone setup - but separate from that it's also used for HRmax calculation - which is going to be used for formula for calorie burn when HR based too.

    But I was referring to the HR zones - not what your HR is actually at.

    Just as you commented - you set custom HR zone more accurate for age, and it didn't change anything.

    That's what I meant.

    You get a avgHR of 140 for a workout, and you'll burn XXX calories.

    You change the HR zones and that won't change - merely where it appears your workout was for zones.

    Also - changing height would be better than age to tweak Fitbit to use better BMR - that way only stride length is effected, and that is a value that can be changed.
    Age changes several other calculations - none of which you have manual control over.
  • BarbieAS
    BarbieAS Posts: 1,414 Member
    heybales wrote: »
    Well, age is used in default zone setup - but separate from that it's also used for HRmax calculation - which is going to be used for formula for calorie burn when HR based too.

    But I was referring to the HR zones - not what your HR is actually at.

    Just as you commented - you set custom HR zone more accurate for age, and it didn't change anything.

    That's what I meant.

    You get a avgHR of 140 for a workout, and you'll burn XXX calories.

    You change the HR zones and that won't change - merely where it appears your workout was for zones.

    Also - changing height would be better than age to tweak Fitbit to use better BMR - that way only stride length is effected, and that is a value that can be changed.
    Age changes several other calculations - none of which you have manual control over.

    Ah, gotcha. Yeah, the specific zone doesn't make a difference, but the number does. We were saying the same thing :).

    What advantages does changing my height give me over changing my age? As I said, I did try height first, and I had it that way for several months, but it doesn't work very well for me because I have to lower my height from 5'5" down to 3'4" to get to the right BMR number (per that super fancy amazing Excel spreadsheet that I think you put together!). I did that for awhile, but I'm not able to adjust my stride at that height (since my actual stride is about 75% of the total height I had entered, it wouldn't allow me to enter that number) which I think was throwing things off even more, and it made my BF% measurements on my Aria scale useless.

    So, I moved to changing my age, just using an online calculator and using trial and error to find an age that matched my actual RMR (actually, I think I would have had to enter something like 135 to get to the right BMR so I ended up using the max age Fitbit would allow [116] and taking a couple inches off my height as well). What else besides calories burned does changing my age affect? I guess demographic stats, but I don't care about those for myself and if Fitbit can't give me a way to effectively adjust my RMR then I guess I don't care about their data integrity either ;). Honestly, the TDEEs that I'm getting over the last 2 weeks are far closer to my "actual" average TDEE based on a few years of food logging vs. weight change than anything Fitbit has ever given me before, so I think I'm on the right track. However, if there are more factors that I'm not considering I'd truly love to hear them, because my only goal is for as accurate data as I can possibly get within the limitations of the program.
  • heybales
    heybales Posts: 18,842 Member
    edited March 2016
    Wow, that's interesting they allow max age 116 - that is hopeful or very forward thinking.

    Height and age are both going to cause problems with the HR-based calorie burns - because BMI is used (so wrong height throws that off), and age and gender is used for deciding where in BMI range and corresponding VO2max you fall (so wrong age throws that off).

    So it's true, if non-HR-based model, I'd say height for sure. Accept I didn't know about limits to the stats they allow you to enter there.

    With HR-based model, I'd be curious to run some numbers to see which could cause worse effect.
    But since HR-based calorie burn is going to be exercise, and that is what, 1 hr of 24 where it applies, so even if it was 50% off, no big deal for the day in total.

    So indeed, seems you've tweaked it right.

    Your tested RMR was really that much lower than calculated by Mifflin or Katch using BF%?
    I'd be a tad concerned with that.
    I will suggest that in the following case, the weight gain was during binge times, which I'll bet she had more than a few of.
    A similar case study was published by Jampolis (2004).
    A 51 year old patient complained of a 15 lb weight gain over the last year, despite beginning a strenuous triathlon and marathon training program (2 hours per day, 5-6 days per week).
    A 3 day diet analysis estimated a daily intake of only 1000-1200 Calories.

    An indirect calorimetry revealed a resting metabolic rate of 950 Calories (28% below predicted for age, height, weight, and gender).

    After medications and medical conditions such as hypothyroidism and diabetes were ruled out, the final diagnosis was over-training and undereating.

    The following treatment was recommended:
    Increase daily dietary intake by approximately 100 Calories per week to a goal of 1500 calories
    32% protein; 35% carbohydrates; 33% fat
    Consume 5-6 small meals per day
    Small amounts of protein with each meal or snack
    Choose high fiber starches
    Select mono- and poly- unsaturated fats
    Restrict consumption of starch with evening meals unless focused around training
    Take daily multi-vitamin and mineral supplement
    Perform whole body isometric resistance training 2 times per week

    After 6 weeks, the patient's resting metabolism increased 35% to 1282 Calories per day (only 2% below predicted).
    The patient also decreases percent fat from 37% to 34%, a loss of 5 lbs of body fat.

    Jampolis MB (2004) Weight Gain - Marathon Runner / Triathlete. Medicine & Science in Sports & Exercise, 36(5) S148.
  • BarbieAS
    BarbieAS Posts: 1,414 Member
    Yeah, all of that is the main reason I moved from the Charge HR to an Alta - I think that HR-based monitoring is too far off for me without the ability for a lot more customization. Even if it's not perfect, I think I'm getting a LOT closer with the settings that I have now than I have at any point in the past with a Fitbit. Time will tell - I'm waiting for a few weeks more weeks of data to be sure.

    And, yes, tested RMR comes in at 1,350. Online calculators range around 1,750, give or take a bit depending on the method.

    Trust me, I'm concerned with it, too. But medically I test perfectly in all respects - thyroid, hormones, etc. And I don't think it's induced by over training or under eating - it's not like I had a period where I lost well and then it tapered off; I've lost about 20lbs (of the ~90-100 total I needed to lose after having kids) in the not quite 4 years I've been on MFP and it's all come in bits and chunks over that time. I might just honestly, truly, be a special snowflake ;).

    Thank you, though, for that study. I haven't seen that. I do tend towards most of the treatment items listed but I could definitely tighten up a bit on a few - it's certainly worth a try.
This discussion has been closed.