FitBit Step Count Review

I tested the step counting accuracy of the FitBit Charge 3, FitBit Charge 2, Xiaomi Mi Band 3 and Spire Stone. Surprisingly, both the FitBits are by far the worst and the Xiaomi Mi Band 3 is the best!
Jump forward:
1:43 Results 10,000 steps test
5:33 Results walk test
6:23 Results activity test
8:14 Single day step graph
Your FitBit is bad at counting steps!
10.000 steps a day, that has become a goal for many of us, and tracking your steps is easier than ever. Now-a-days, Step counters, also called pedometers, are standard features of most smart-phones and wearables. However, as a user of several different step counters, I really started doubting their accuracy. At the end of the day, I often see big differences between the step counters I use. In this video I analyze the accuracy of several step counters, and see what they do right and wrong. In short, I find that all step counters are good at counting steps when you are actually walking, however, some step counters count many false steps when they are not supposed to.
Icons used from www.thenounproject.com:
“Bus” by i cons
“Car” by shashank singh
“dumbell” by Arif fauzi hakim
“Train” by Gergor Cresnar
“ergometer” by Falco Bosser
“Cycling” by Adrian Coquet
“Walking” by Irine Hoffman

[@AleksiJoensuu]

My girlfriend's Fitbit Inspire HR congratulated her on a good swim after she was done drawing and cutting fabrics for some clothing she was making :D

[@TheSongBirdRainStar]

🤣

[@nightsgrow6575]

@Fourstar mine always gives me swimming when I’m vacuuming and when I’m folding laundry 😂

[@RichardWerbin]

I have a test suggestion. I am 78 years old. When walking slowly the Charge 2 seems to count steps accurately. But, the Versa 2 counts half the number of steps counted by the charge 2. If I walk briskly, the steps counted by Versa 2 and Charge 2 are approximately the same. Can you recruit and older, slow walker and test step count accuracy?

[@StatesideBeast]

Fitbit Versa 2 - I walk 100 steps I will get 128-158. I take 8 steps to the bathroom and I get 174 steps... I have contacted support. I have reformatted 5 times! I think I'm done with Fitbit.

[@retrieverlover8228]

Check out the FITBIT SUPPORT Twitter feed. They put out a statement just yesterday about the firmware.

[@richlaue]

It all comes down to the arm movements.

[@TheQuantifiedScientist]

Sorry to hear that. My experience with FitBit support has been really good, hopefully you got it fixed!

[@viksaggu9085]

Rich Laue yup

[@hanswurst1541]

get the Mi Band 5 its better

[@gwerjio83]

Amazing. Could you do the same with the latest trackers as of end of 2022 please? I also started testing a couple of bands after suspecting that my new Fitbit Charge 5 is overcounting steps but my calculations are nowhere near as sophisticated as yours. I also noticed that all the bands/trackers are very accurate when walking outdoors, but some overcounting when at home.

[@GianlucaTruda]

Awesome experiment, Rob!

[@TheQuantifiedScientist]

Thanks!

[@charlottelumley8196]

Really cool video! Just wondering whether when cycling/home training you were using the built in functions on Fitbit to register those exercises as cycling exercises & steps were counted regardless, or were the functions not used? Also you seem to be wearing the fitbits on different wrists - did you alternate wrists to account for having a dominant hand? Could explain why FBC3 counted so many more steps than FBC2 to some extent.

[@TheQuantifiedScientist]

Hi Charlotte! Yes, those functions were used, so both FitBits knew what sport I was doing. I actively told it I was training and cycling outside was detected automatically. I did wear them on different wrists most of the time, so that could indeed explain it. I had thought about it, but it was just not comfortable long term on a single wrist. I did do a check now with the FB charge 2 and Charge 4 on the same wrist for a few weeks, I still need to check the data :-). Thanks for you good suggestions Charlotte!

[@AleksiJoensuu]

@@TheQuantifiedScientist Did you change the setting from dominant to non dominant hand in the Fitbit app when you changed wrists? It would be cool if that were an easier, single press function in the fitbit app especially for those of us who still use old-timey metal-chunkey tickey-tockey watches when we are not sportsing about.

[@cfrost87]

My fitbit charge 3 usually counts a couple hundred steps while I'm sleeping. I guess I move around enough to make it think I'm walking. So my step count is probably overestimated by 300-500 steps a day, but I don't expect it to be perfect. As long as I can get a good estimate, I'm ok with that. I think trackers like these can be helpful in getting people to be more active. Mine even has a setting where you can have it remind you to move every hour, if you've been sitting and not gotten up to move around.

[@TheQuantifiedScientist]

Hi Chrystal, it indeed depends on your goal! For you it seems to work well :-D . I would just personally hope that the steps were accurate, or otherwise they would use a different metric, like "Energy Spent" or something else as the main indicator of activity, instead of steps. Have a great day!

[@knightofhope2479]

@@TheQuantifiedScientist best fit fitbit you recommend for steps?

[@ChantelleDinkel]

Interesting content! Quite detailed tracking! Thank you! Did you by chance also keep track of floors and how accurate that was? I'm really interested in that as I like to hike and maybe will go on a long-distance multi-month hike and thought elevation gain would be interesting to keep track of, not just steps, though that's interesting as well. :)

[@TheQuantifiedScientist]

Hi Chantelle, unfortunately I was not able to track that. The FitBit Charge 2 might actually have an altimeter built in (a quick internet search was not 100% conclusive), so that should be able to pretty accurately track your elevation. If you have another Fitbit it might have GPS in it, which would also be able to accurately track your elevation. However, I never actually tested it. If you have any feeling about the accuracy after going on your hike, I'd be interested to hear about your experience :-) . Have fun on your trip and good luck!

[@cfrost87]

I have a charge 3, and it keeps track of floors, but the floor tracker is so inaccurate. It counts elevation as stairs even when you don't go up or down stairs, and elevators as stairs.

[@zapiekank]

I am considering switching from Fitbit to something else because, besides the fact that I've already had 2 replacements, steps and sleep seem to have gone mental. I'm still on Versa 1, first one died for no reason and the 2nd the screen popped out, again for no reason, no impact or anything causing it. Now on my 3rd Versa, sleep is all over the place, half of days of the week doesn't even pick it up and also it counts way too many steps. This week I bought an Amazfit GTS to try see if it's better. I've been wearing the two all day, one on each wrist, and at the time of writing this (10.30pm UK, also bear in mind I've not been on a proper walk because weather wasn't too pleasant) Amazfit has counted 7513 steps and Versa says 13805 steps. Yes, Versa was on my right wrist, so when I was cleaning the kitchen cupboards it counted steps. But 6000 steps from cleaning my kitchen is a tad much? Now, I don't know how accurate Amazfit is, but in my less than scientific opinion, Versa seems to most certainly be taking the piss here. Thank you for your dedication, time and effort reviewing these bands! You're awesome!

[@MagisterVeritas]

Very nice work. Subscribed within the first two videos of yours I watched. Maybe also invest in some foot pods (Garmin has them for example) which you tie on your laces. I am curious if those track steps better. For example I have a Garmin Fenix 3HR and at some point I saw it tracking steps while riding the tram. Good luck!

[@TheQuantifiedScientist]

Thanks Alex! I'd never heard of Foodpods, definitively worth checking out. Thanks for subscribing!

[@jasonengel2374]

Does your phone track your activity too? It would be interesting to see if your phone was more accurate. My hypothesis being that because your phone has GPS it can calculate velocity etc. As I am an Android user I have noticed that google is pretty good at tracking my daily activities and knowing when I am taking a cycle and stopping off at a shop or park - even if it's only a very short stop. Thanks again for all these great videos! I only stumbled across you today but I've watched and thumbed them all :-) Cheers

[@TheQuantifiedScientist]

Thanks so much Jason!! I have an iPhone, but the problem is that it is not always in my pocket. Around the house it will often be on the table thereby missing quite some steps. I must admit i have never checked, but it might really vary from person to person depending on how they take their phone with them😊. Thanks for the idea!! Have a great day

[@hectornonayurbusiness2631]

Why is this a problem? When you are on your trainer or cycling you are still active. I would think they are equivalent to steps.

[@georgev.]

Do you have data about the battery duration? Great video thanks.

[@sidbrun_]

Could you do this test using different phones? Since almost everybody tracks their steps using their phone, would be interesting to see which are the most accurate.

[@yozy4996]

Excellent Job...Thank you for sharing your time and effort..

[@TheQuantifiedScientist]

My pleasure!

[@danstarr9895]

Amazing work. Great job collecting all this data.

[@AleksiJoensuu]

Probably most Fitbit owners will have had one of those days when you haven't gotten outside of your apartment and your Fitbit is showing hundreds of steps. Where the heck did I walk to? :D

[@anonymouswhite352]

Only hundreds would be you just walking around your house

[@aprilsmiley7025]

My issue with my fitbit versa is , I go to bed at night with the steps accurate , but in the morning , it shows that I have like 31,000 steps or this morning it's showing 21,000 steps. I try to resend and resend and it doesn't fix it until usually the next day

[@boimesa8190]

Which is the best sleep tracker??

[@retrieverlover8228]

So helpful because I bike a lot (outside when the weather is nice; and now inside because it's horrible!) and my steps are way off (I have a Charge 2; glad I didn't get the 3 now!). Not sure how to resolve this other than taking it off or guesstimating how many extra steps I am getting per 10 minutes of riding. It is really frustrating because steps and cycling (mileage) are both important to me and I have my Fitbit synced with My Fitness Pal app (I have been inputting cycle data manually which means I have basically doubled up the activity) and Map My Ride app (outdoors only obviously). UGH! If you have a different suggestion for maintaining some level of accuracy other than trying to remember to take it off when I ride, I would love to hear it! 😊 Thank you for this video. Super interesting and you took a lot of time to do all this!

[@TheQuantifiedScientist]

L S thanks for your reply, and I am happy to hear that the video was of use to you!

[@TheQuantifiedScientist]

L S , it is tricky to do it any better with the Forgot. Your FitBit should recognize automatically when you go cycling though, you could subtract the steps from those activities from your total. Be aware you can set the minimum time before it will register it (i think you can go as low as 10 mins). Let me know if this helps

[@deadbones23]

Very cool analysis 👍

[@marcogtturbo]

Impressive job. Just what I was looking for. Im about to find a way to track my child steps who's affected by miopathy while is under treatment with a drug .he's nearly 4 yo and 1m tall. But the issue is that he's strolling for long distances. I also need a good access to raw data. I'm thinking about Fitbit inspire hr but I'm still not convinced. What could be your suggestions ? Thank you

[@TheQuantifiedScientist]

Hi! Thanks for commenting. So in this case i find it difficult to give clear advice, since this is a very specific case you are describing. what you could do is buy a device and do a few small test. Take a situation where you should not get steps, and see if you do. Next test a situation where you should get steps, and see if you get them. As to which device you should get, it also depends what your child finds comfortable perhaps. I cannot see into your personal financial situation, but FitBit allows you to download all your data in .json format. So you could try it and perhaps return it if it does not meet your standard? If you just want to check total steps at the end of the day you could also buy a cheap " traditional " step counter that is worn on the belt. Good luck with your situation and good health!

[@marcogtturbo]

@@TheQuantifiedScientist I'll give fitbit inspire hr a try, I can easily find it used so it worth to try. My idea is to create a dataset with several additional variables in order to create homogeneous clusters of data and identify trend of improvement (hopefully) thank you

[@TheQuantifiedScientist]

@@marcogtturbo Good luck, and let us know if it works out!

[@bascal133]

Can you look into the whoop strap?

[@TheQuantifiedScientist]

Hi Bacall, thanks for your interest in my video and for commenting! I would love to look at the whoop strap, but unfortunately it is too expensive for me at this moment. It seems to be a monthly subscription service, which is a little out of my price class right now. But if you have any experience with the whoop strap, please share it!

[@tinybandit5203]

what's the name of the iphone app you mention ?

[@TheQuantifiedScientist]

Tiny Bandit, thanks for your question. Do you mean the location tracking app? The app I use for that is "Arc App".

[@demiangarcia-monroy2564]

Mi Band 5 stops recording my steps for a bit and then it catches up but most of the time it only logs 50% of the steps than I actually did. I find Mi Band 5 more accurate when walking for 100 steps or more at a good pace. if you are walking intermittently and at a slower pace, it has been extremely inaccurate for me.

[@GrayWithMe]

How about the step counter on the Apple Watch ?

[@TheQuantifiedScientist]

Hi! Thanks for your comment :-). That is something I still need to test, so stay tuned :-)

[@george6252]

Wear your Fitbit underneath your sock above your ankle. That's where I rotate weekly my charge 4 / charge 5.

[@fomalhauto]

I tend to move both my legs and arms rapidly. I will tend to move my arms like I am dancing when I am walking to dance music that I am listening to. I also did rapid arm movements that I called "thigh drummers" (I hit my thighs rapidly) as a type of cardio when I am riding the bus.

[@D3ft0ne]

Yep... Switched to garmin, and found out.. All the 10k's before with fit bit were sweet little lies.

[@Pug318]

I just purchased Fitbit charger 5 and it’s counting 3-4 thousand extra steps everyday just by moving my hands - I work as barber and as I am cutting it’s sayings I’m walking. I tried to reduce sensitivity as much as possible it did help, but it still count steps while I’m standing 🤦🏻‍♂️

[@monideun]

Would be cool to see how a phone's step counter works in comparison!

[@richlaue]

Pprobably a little better.

[@TheQuantifiedScientist]

Indeed, that would be very cool! I suspect the accuracy is highly dependent on two things: (1) Where do you keep your phone [in a bag, or in your pants pocket], and (2) do you carry it on you at all times [or is on on the table at home quite a lot]. Thanks for your comment Simona!

[@TebTengri]

My Ionic kept giving me steps when driving as I assume it'd count steps when I'm moving that hand on the wheel.

[@TebTengri]

@Stelios The Greek I'm actually tempted to put a sweatband on that wrist, not under the fitbit but between it and where the ulna pokes out to keep it from sliding up and down my arm.

[@kasey5602]

My Fitbit Charge 4 counts my steps while I cycle, it drives me crazy

[@Donilan]

Could you please test some more Withings devices?

[@TheQuantifiedScientist]

Hi Tristan! Hopefully I can :-) . I did a review on their thermometer: kzbin.info/www/bejne/bZ3XpHmdjNOKfbc ; and their sleeping mat: kzbin.info/www/bejne/gJDRZomlgdp_p5Y , which you might have seen. I also have their scale and their BP monitor that I've been using for 3 years. Do you have any experience with their products?

[@_Pumpkin_Seed_]

I wish you also tested an apple watch vs fitbit. Shame this video does not have more views. It doesn't matter which apple watch, series 3, series 4 or series 5. Just to see if apple can do it or not

[@TheQuantifiedScientist]

Hi Chewy, yes for sure. I am thinking about buying the new Apple watch once it comes out. I will keep you posted! (It should also have build in sleep tracking). Have a great day!

[@viksaggu9085]

The steps are also based on your stride length to. Lots of factors

[@TheQuantifiedScientist]

Hi Vik! Thanks for commenting! For sure, for some devices that would play a role, especially GPS based ones. However, most of these will be base on the accelerometer data I suspect, which I think would be less influenced by stride length. They would just search or some rhythmic pattern to count steps. But I might be missing something, what do you think? Have a great day!

[@richlaue]

Do you really expect a device that looks at arm motions to accurately count the movement of the feet?

[@TheQuantifiedScientist]

Hi Rich, thanks for your question. Not an easy one to answer, but I will give it a shot. I tested 4 devices: 3 wrist-worn and one on the strap of your pants. Interestingly, the most accurate device was one of the wrist-worn devices, however, the least accurate was also wrist worn. I therefore suspect that with the right software wrist-word step counters can be pretty accurate.

[@cfrost87]

If you are moving your arms with the normal swing that happens during walking, it will fairly accurate.

[@fomalhauto]

I tend to move both my legs and arms rapidly. I will tend to move my arms like I am dancing when I am walking to dance music that I am listening to. I also did rapid arm movements that I called "thigh drummers" (I hit my thighs rapidly) as a type of cardio when I am riding the bus.

[@chronometa]

I personally would count cycling as steps. It's practically the same

[@joshuaberry4128]

Your hands were held in front of you using the counter. Have you done it with hands swinging in a normal human gate?

[@rahimyazid]

Sleep test with these bands?

[@TheQuantifiedScientist]

Hi Rahim! I did test the fitbits: kzbin.info/www/bejne/hIq1omeXnbV9aMU and kzbin.info/www/bejne/i3LGkqZ8g9-Iopo

[@TheQuantifiedScientist]

I will over time add other bands as well :-) . Have a great day!

[@rudys0951]

Would you be so kind as to make one for the Apple watch as well please 🙏🏻

[@TheQuantifiedScientist]

Hi RudysFarinas, thanks for the comment! When I get the opportunity, I definitely will :-D. Have a great day!

[@athletejmv]

What this person said. I was using Fitbit and it was way too generous giving steps... HR sensor too being +40BPM more than actual. Apple Watch has been spot on... trying to get away from the Apple Watch but it has been the most accurate for my uses.

[@TheQuantifiedScientist]

@@athletejmv thanks for letting us know! If I ever can test it, I will :-D

[@saotul]

What about an update video

[@joseph19761]

My Fitbit 4 counts horrible, I’m a truck driver and it’s counting a lot while I’m driving.... I called Fitbit the only thing they could do , they said is to take the watch off while I’m driving . or delete the hours on the app that I’m driving.... which my Apple I phone counted more accurate in my pocket...👎🏼for fitbit

[@421prateek]

perfect

[@LOTPOR0402]

Step counter is irrelevant anyway ,distance and speed are more important .You would think a watch with GPS would do that pretty accurate

[@MandyGood]

Omgosh I had no clue :(

[@TheQuantifiedScientist]

Happy to help :-) . Have a great day!

[@lyubomirhristov418]

Why always when I shake my hand sitting down my fitbit4 count steps it’s shame

[@TheQuantifiedScientist]

Hi! Yes, all these step counters are based on accelerometers, so quite prone to errors if the algorithm is sensitive to other movements :-|. Have a great day!

[@lyubomirhristov418]

The Quantified Scientist cheers thanks

[@andrewhwang2910]

Wow, you manually counted *forty thousand steps* (ten thousand for each of the four devices) to test sensitivity? And somehow that was by far the least impressive part of this analysis! Your reviews blow everything else out of the water, thank you for sharing Rob.

[@aguyfrommanitoba4981]

Your summary is right and wrong... they count aerobic exercise on a scale of steps..... so yes they are counting steps for your home trainer.... I don't get why you are so keen to count distance? The fitbit is about overall activity buddy

[@boibluex12]

dude make more videos wtf??

[@TheQuantifiedScientist]

Haha, thanks William !! I will do my best, thanks for your support :-). Have a wonderful day!

[@thisislorn]

the fitbit counts an obscene amount of fake steps while driving a motorcycle lmao

[@hanswurst1541]

he wrote a computer program, by the way

[@death.noneexistentchannel5797]

Hey stepH

[@bstal26]

Just get a Pokéwalker

[@chronometa]

So in this case less technology is better.

[@theneverending9319]

I wouldn't trust the xiaomi mi. The company is known for data collecting. Also the watch comes in Chinese not English

[@fomalhauto]

What do you mean the watch comes in Chinese not English?

[@aratosm]

Saying Xiaomi is most accurate after earlier stating you wore it for a month compared to 18 months for all the other ones is so so so misleading.

[@TheQuantifiedScientist]

Hi Aratosm, thanks for your feedback! I appreciate your input, but in this case I have a different opinion, let me try to explain why. My habits on most days look more or less the same: I get up, exercise on my home-trainer, cycle to work, work, cycle back and on 2 out of 5 nights I exercise in the evening. This means that wearing a device for a month will give me a pretty representative sample of how it performs. I have tried analysing the FitBit results on shorter timespans, and this did not change results significantly. Additionally, if you look at the example plots I provide for single days, you can see the exact same patterns re-occuring. These confirm my general presentation of the findings. Also, let me state I have no interest in making one device look better than the other, I always try to objectively judge the quality of the devices. So overall, I think that having data for one month in sufficient in this case, and not misleading or biasing the results. What is your reasoning that it is? Have a great day!

[@charlottelumley8196]

Really cool video! Just wondering whether when cycling/home training you were using the built in functions on Fitbit to register those exercises as cycling exercises & steps were counted regardless, or were the functions not used? Also you seem to be wearing the fitbits on different wrists - did you alternate wrists to account for having a dominant hand? Could explain why FBC3 counted so many more steps than FBC2 to some extent.

[@TheQuantifiedScientist]

Hi Charlotte! Very good questions! So, I thought I'd responded already, but apparently, something went wrong. To answer your question: yes, I always tell the fitbit that I am on my home trainer to track my activity, and my cycling is picked up on automatically as an activity. About the wrists: yes you are right. For a while I had them on the same wrist, but this was not comfortable long term, so I decided against it in the end. I could have alternated, but this was practically not easy to do. So you are right, this could be part of the reason. However, I don't know if the movements of my right arm are drastically different during cycling and when I am on my hometrainer. What do you think?

[@charlottelumley8196]

The Quantified Scientist Hey! I actually did get your previous reply and thought I had “liked” it.. how strange! No I can’t imagine the wrist would make too much difference on the bike. I found mine tracks significantly more steps when I wear it on my dominant hand for activities like ironing, cooking and cleaning though. Not enough to count for the whole discrepancy I wouldn’t have thought, just worth a mention!

[@TheQuantifiedScientist]

@@charlottelumley8196 Hi Charlotte. Yeah, for cooking, ironing, etc. it would make a difference for sure. Thanks for the great comment! Have a great day :-)