Fitness trackers are terrible at counting calories, says Stanford study
RobBasss
Posts: 65 Member
Replies
-
"So Ashley's team evaluated 7 devices -- the Apple Watch, Basis Peak, Fitbit Surge, Microsoft Band, Mio Alpha 2, PulseOn and Samsung Gear S2 -- with a group of 60 volunteers using rather than using treadmills and stationary cycles in the lab. The researchers compared the devices to FDA-approved gold standards, rather than to each other. "
I think they left something out of that sentence. Using what instead of dreadmills and hamster wheels?7 -
How did they estimate intake and weight change? That plays just as much of a role as anything.
Your food logging, type of exercise and the variance between your actual and predicted BMR as well as the total amount of exercise you do in a day all play a role in final accuracy.
That said...shrug re studies. Mine is accurate to 4% and that's good enough for my needs!
6 -
Were these devices with heart rate monitors or are these purely based on movement?1
-
How did they estimate intake and weight change? That plays just as much of a role as anything.
Your food logging, type of exercise and the variance between your actual and predicted BMR as well as the total amount of exercise you do in a day all play a role in final accuracy.
That said...shrug re studies. Mine is accurate to 4% and that's good enough for my needs!
I linked to a better article, explains more, anyhow I am suprised by the findings, I use my Vivosmart 3 to guage my daily burn and based on my pretty consisitent 1 or so pound a week loss, it can't be too far off.0 -
It's a rough estimate... study discovers...7
-
So here is the thing.
They measured a specific exercise burn.
And the trackers and any other form of estimation are unreliable. Even measuring the oxygen is not 100% reliable.
But my tracker is useful not because it measures a specific exercise accurately but because it measures my daily tdee impartially, automatically, without me having to overthink it, and with sufficient accuracy over time for me to make eating decisions with 99% confidence. Because over time I know that it is at best off by 4% when it comes to the whole day.
Do I have to engage SOME brain power? Yes, when I changed trackers halfway through the day and when I changed time zones the tracker was substantially off and I had to figure out on my own that the extra 1000 Cal were not really there.
By the way note that my 4% includes those incidents...
Anyway. Yes. They are an estimate that you combine with your results and will only be as accurate as you make them (see accuracy of most peoples' food logging)10 -
The MOST accurate device (the apple watch) was still off by 27%!!!! The worst was off by 93%. They also were even more inaccurate for anybody with dark skin and those with high BMIs.
Bottom line: these things are very good at tracking heart rate, BUT these things are basically worthless in terms of estimating how many calories you burn through exercise. There are just too many variables involved to create one algorithm that will be accurate for everybody.0 -
BeJeanne11 wrote: »Were these devices with heart rate monitors or are these purely based on movement?
That's answered in the link in the first post of this thread.
1 -
But my tracker is useful not because it measures a specific exercise accurately but because it measures my daily tdee impartially, automatically, without me having to overthink it, and with sufficient accuracy over time for me to make eating decisions with 99% confidence. Because over time I know that it is at best off by 4% when it comes to the whole day.
For context, this relies a great deal on your behavior, which sounds remarkably consistent from one day to the next, and may not work for many people.MoiAussi93 wrote: »Bottom line: these things are very good at tracking heart rate, BUT these things are basically worthless in terms of estimating how many calories you burn through exercise. There are just too many variables involved to create one algorithm that will be accurate for everybody.
Algorithm, sure. On the other hand, a direct force power meter will never be off by more than 5 % for bike calories, and generally most computers will put you in the middle of that range so it might be over or under by as much as 2.5 %. Sadly this technology isn't available for things like walking or skiing.1 -
Here's a another article about it from NPR. http://www.npr.org/sections/health-shots/2017/05/24/529839681/fitness-trackers-good-at-measuring-heart-rate-not-so-good-at-measuring-calories BTW, I have a good friend who swears by her fitbit, records her food, and doesn't understand why she isn't losing the weight she expects to be...this may explain it at least to some degree...rest assured I am emailing this to her. I think the quote below is VERY relevant to the way that many people (at least on this site, and the friend I mentioned above) try to use these things.
"It's just human nature. People are checking these inaccurate counts and they think they've earned a muffin or earned some ice cream and they're sabotaging their weight-loss program."
Church points to a study last year which found participants in a weight-loss program who also wore fitness trackers actually lost less weight than participants who didn't wear the trackers. "It's an instance of no information is probably better than having bad information," he says.2 -
MoiAussi93 wrote: »The MOST accurate device (the apple watch) was still off by 27%!!!! The worst was off by 93%. They also were even more inaccurate for anybody with dark skin and those with high BMIs.
Bottom line: these things are very good at tracking heart rate, BUT these things are basically worthless in terms of estimating how many calories you burn through exercise. There are just too many variables involved to create one algorithm that will be accurate for everybody.
no they aren't...that's what is funny.
The can measure movement tho and apply a calculation just like everything else does...
you learn as you use them what to do and how you react...it's pretty easy if you want it to be.
As well most people use them as motivation to move for fitness anyway and if that's all they do it's worth the funds.4 -
MoiAussi93 wrote: »The MOST accurate device (the apple watch) was still off by 27%!!!! The worst was off by 93%. They also were even more inaccurate for anybody with dark skin and those with high BMIs.
Bottom line: these things are very good at tracking heart rate, BUT these things are basically worthless in terms of estimating how many calories you burn through exercise. There are just too many variables involved to create one algorithm that will be accurate for everybody.
no they aren't...that's what is funny.
The can measure movement tho and apply a calculation just like everything else does...
you learn as you use them what to do and how you react...it's pretty easy if you want it to be.
As well most people use them as motivation to move for fitness anyway and if that's all they do it's worth the funds.
However they measure/estimate it, most of the devices are only about 5% off on heart rate. For a consumer oriented device (i.e., NOT a medical device meant for clinical use) that is VERY good. It means if you use it to try to keep your heart rate in a certain zone during exercise, this will work very well. If you are trying to track resting heart rate, the data is pretty reliable. What you see accurately reflects reality.
Calories are a very different matter unfortunately.
If wearing a fitness tracker motivates you, great, but many people don't need to pay for that motivation. I am motivated to move without one. But if people do pay for one, they need to understand the limitations and not assume those calorie counts are accurate, since the data says they are not.
1 -
I have a Fitbit Charge 2 and use it for 2 reasons: to tell time and count steps. The number for steps is for just one reason: to ensure that I'm at least being active throughout the day. If the steps are low, I move more.
A.C.E. Certified Personal and Group Fitness Trainer
IDEA Fitness member
Kickboxing Certified Instructor
Been in fitness for 30 years and have studied kinesiology and nutrition
13 -
MoiAussi93 wrote: »However they measure/estimate it, most of the devices are only about 5% off on heart rate.
I did a run the other day. My wHRM read 120 bpm. I was above my threshold which is 165 bpm. This is with a high end tracker which has only been widely available for purchase for about 6 weeks.1 -
MoiAussi93 wrote: »MoiAussi93 wrote: »The MOST accurate device (the apple watch) was still off by 27%!!!! The worst was off by 93%. They also were even more inaccurate for anybody with dark skin and those with high BMIs.
Bottom line: these things are very good at tracking heart rate, BUT these things are basically worthless in terms of estimating how many calories you burn through exercise. There are just too many variables involved to create one algorithm that will be accurate for everybody.
no they aren't...that's what is funny.
The can measure movement tho and apply a calculation just like everything else does...
you learn as you use them what to do and how you react...it's pretty easy if you want it to be.
As well most people use them as motivation to move for fitness anyway and if that's all they do it's worth the funds.
However they measure/estimate it, most of the devices are only about 5% off on heart rate. For a consumer oriented device (i.e., NOT a medical device meant for clinical use) that is VERY good. It means if you use it to try to keep your heart rate in a certain zone during exercise, this will work very well. If you are trying to track resting heart rate, the data is pretty reliable. What you see accurately reflects reality.
Calories are a very different matter unfortunately.
If wearing a fitness tracker motivates you, great, but many people don't need to pay for that motivation. I am motivated to move without one. But if people do pay for one, they need to understand the limitations and not assume those calorie counts are accurate, since the data says they are not.
I think that like most things...aka exercise machines or other ways of getting calorie burns it's an estimate and most get that....for example...here on MFP it is often said only eat back 50% of the exercise calories as they are over estimated...but they weren't for me. I could eat them all.
The calculation is an average and will work for the average....imo. As well user error and setting them up correctly has a big impact as well. For example if you say you are left handed and are wearing it on your left hand but really it's on your right your readings will be off.
People have to use "common sense" when looking at this stuff...for example do I believe a 3mile run burns 300 calories...probably not but it's pretty close....3 -
NorthCascades wrote: »
For context, this relies a great deal on your behavior, which sounds remarkably consistent from one day to the next, and may not work for many people.But my tracker is useful not because it measures a specific exercise accurately but because it measures my daily tdee impartially, automatically, without me having to overthink it, and with sufficient accuracy over time for me to make eating decisions with 99% confidence. Because over time I know that it is at best off by 4% when it comes to the whole day.
No, what it relies on is the accurate measurement of food intake and weight change over time in conjunction with consistent tracking by the tracker. It IS dependent on performing activities that "register" with the trackers. For example neither swimmers, nor people whose movement is not detected by their tracker need apply...
Beyond that daily behaviour is somewhat irrelevant and my TDEE varies from 2100 Cal to over 4000 Cal in a day.
As to your friend.... they are missing the part where they have to adjust based on real world results.
3 -
I use my fitbit to remind me to move more when I'm being lazy so that I can have a tidy little NEAT at the end of the day. I don't personally eat differently from day to day (unless I'm particularly hungry or have plans to have a large meal with family etc) My fitbit reports an average burn of 2200-2400 calories for my daily activity, as do most online TDEE calculators. These are just 'best guesses' and what you should actually use is your own data. Is it necessary for me to eat 1200 based on the estimate? No. Will I eat 2K and expect to lose? Probably not either but it keeps me in a comfortable middle ground with realistic expectations.
That said, my rate of loss measured against my intake closely matches the estimated TDEE my FitBit reports, so there's that.0 -
Didn't bother reading the article. I've never used my Fitbit for anything other than to count steps and floors. And occasionally I will use the sleep tracking functionality.0
-
NorthCascades wrote: »
For context, this relies a great deal on your behavior, which sounds remarkably consistent from one day to the next, and may not work for many people.But my tracker is useful not because it measures a specific exercise accurately but because it measures my daily tdee impartially, automatically, without me having to overthink it, and with sufficient accuracy over time for me to make eating decisions with 99% confidence. Because over time I know that it is at best off by 4% when it comes to the whole day.No, what it relies on is the accurate measurement of food intake and weight change over time in conjunction with consistent tracking by the tracker.
I must have done a bad job of explaining my objection if that's your answer. You're telling us your Fitbit is accurate to within 4 % of the gods' honest truth, but that isn't widely applicable. I don't know what you do for exercise but to use your example, the accuracy of your Fitbit depends on you not swimming. People who swim, and expect 96 % accuracy from their Fitbit, will be disappointed.As to your friend.... they are missing the part where they have to adjust based on real world results.
In my opinion, either a measuring device is accurate to some specification (no more than 4 % error was claimed) and its readings are trustworthy within spec, or it's not accurate and must be adjusted. It's like saying my thermometer is perfectly accurate, I just ignore it when it tells me it's 90 F outside if I can see it's snowing. At this point the Fitbit isn't what's accurate, it's the real world results that are accurate.0 -
I like to see my heart rate while exercising so I will know how hard I am working--I have thought, "Man, I'm really pushing it," looked down at my watch and saw my HR was 135, and found a lot more fuel in the tank. And for running, my Garmin helps me figure out how fast to go. (Again, if I don't look at the number, I tend to go slower.)
As for calories burned, I have never been able to eat back exercise calories, anyway. I simply gain weight when I do that. I might eat an extra 100-200 on days when I have a really rigorous workout, but it's not like I ever can eat an extra 800 calories because I ran 8 miles. My body is used to running 8 miles and doesn't seem to shed anything other than sweat from that experience!2 -
My fitbit charge 2 seems to do ok in keeping my eating in line with mfp synced. Lost 90 lbs and I eat whatever it says I can most of the time. Mfp set at sedentary and the charge does the rest, keep it simple.
2 -
NorthCascades wrote: »NorthCascades wrote: »
For context, this relies a great deal on your behavior, which sounds remarkably consistent from one day to the next, and may not work for many people.But my tracker is useful not because it measures a specific exercise accurately but because it measures my daily tdee impartially, automatically, without me having to overthink it, and with sufficient accuracy over time for me to make eating decisions with 99% confidence. Because over time I know that it is at best off by 4% when it comes to the whole day.No, what it relies on is the accurate measurement of food intake and weight change over time in conjunction with consistent tracking by the tracker.
I must have done a bad job of explaining my objection if that's your answer. You're telling us your Fitbit is accurate to within 4 % of the gods' honest truth, but that isn't widely applicable. I don't know what you do for exercise but to use your example, the accuracy of your Fitbit depends on you not swimming. People who swim, and expect 96 % accuracy from their Fitbit, will be disappointed.As to your friend.... they are missing the part where they have to adjust based on real world results.
In my opinion, either a measuring device is accurate to some specification (no more than 4 % error was claimed) and its readings are trustworthy within spec, or it's not accurate and must be adjusted. It's like saying my thermometer is perfectly accurate, I just ignore it when it tells me it's 90 F outside if I can see it's snowing. At this point the Fitbit isn't what's accurate, it's the real world results that are accurate.
Let's try this again:
Fitness trackers are not accurate. Not within 5% or 10% without any adjustment for a lot of people. I am willing to grant you that.
I will assume that you are willing to grant me that calorie counting is also not accurate. Not within 5% or 10% without any adjustment for a lot of people.
So why have you been hanging around a calorie counting site for the past year and a bit?
Did you somehow manage to find calorie counting useful in spite of the lack of absolute accuracy?6 -
Are those all wrist models? I'd be interested to see a study on a Fitbit One or the Zip. I've always wondered how a wrist device could even record steps accurately, much less calories.3
-
I only use my Fitbit for steps and just to get an "estimate" on calories burned. Not surprised by the study or any other studies I've seen. I've never relied on the machines at the gym for accuracy either because those are always off0
-
Doesn't matter if they're perfectly accurate, aa long as they are consistent. You can adjust eating (like thre common "eat 50% of your exercise calories back" advice) based on your experience if the tracking is consistent.6
-
Are those all wrist models? I'd be interested to see a study on a Fitbit One or the Zip. I've always wondered how a wrist device could even record steps accurately, much less calories.
I wear both a zip and a charge 2 and by the end of the day my steps are usually within a few hundred and calories burned are within 1 to 200 with the charge 2 being a the higher probably because of the hr. I feel that is plenty adequate enough to continue the life journey. Too many of us over think all this.
4 -
Meh, mine is accurate. Of course nothing is going to be perfect but a good portion of fitness tracker inaccuracies are from user error.0
-
Most of the time I find mine to be pretty spot on with my TDEE here on MFP. There are some random days that I end up with a high number of active minutes and I know for a fact I didn't have that many active minutes. I mainly use mine for steps and rely on my Polar HR chest strap for a more accurate reading during my exercise.0
-
PowerliftingMom wrote: »I only use my Fitbit for steps and just to get an "estimate" on calories burned. Not surprised by the study or any other studies I've seen. I've never relied on the machines at the gym for accuracy either because those are always off
This is pretty much true for me, too.
I got my Charge 2 in March, at which point I was starting to reverse diet back to maintenance. I only eat back a fraction of the exercise calories/adjustment but as pp's pointed out, most of us did that with other calorie burn estimates for exercise, e.g. MFP's or the machines at the gym.
MFP estimated my maintenance at 1810 cals/day as lightly active. I think Fitbit is estimating I burn an average of 2300 cals/day but my true maintenance seems to be closer to 1700-1900/day depending on activity level.
0 -
mikelombardo078 wrote: »Are those all wrist models? I'd be interested to see a study on a Fitbit One or the Zip. I've always wondered how a wrist device could even record steps accurately, much less calories.
I wear both a zip and a charge 2 and by the end of the day my steps are usually within a few hundred and calories burned are within 1 to 200 with the charge 2 being a the higher probably because of the hr. I feel that is plenty adequate enough to continue the life journey. Too many of us over think all this.
Just wondering...do you have both synced to a single Fitbit account? Or do you use 2 accounts?0
This discussion has been closed.
Categories
- All Categories
- 1.4M Health, Wellness and Goals
- 393.4K Introduce Yourself
- 43.8K Getting Started
- 260.2K Health and Weight Loss
- 175.9K Food and Nutrition
- 47.4K Recipes
- 232.5K Fitness and Exercise
- 426 Sleep, Mindfulness and Overall Wellness
- 6.5K Goal: Maintaining Weight
- 8.5K Goal: Gaining Weight and Body Building
- 153K Motivation and Support
- 8K Challenges
- 1.3K Debate Club
- 96.3K Chit-Chat
- 2.5K Fun and Games
- 3.7K MyFitnessPal Information
- 24 News and Announcements
- 1.1K Feature Suggestions and Ideas
- 2.6K MyFitnessPal Tech Support Questions