Apple Watch Step Count vs. Fitbit.

[BSG75]

I use both my Apple Watch and Fitbit One to count steps. What I've noticed is that my Fitbit is much more accurate than my Apple Watch. I've calibrated my Apple Watch using the workout app on my watch and walking with my iPhone for 20 minutes. In fact, I've held the phone in my left hand (I wear my watch on my left wrist) during my entire outdoor walk. I've done this numerous times. What I've found is that the Fitbit does a much better job of counting steps.

Today, I did an experiment where I started wearing my Fitbit at the same time I started wearing my Apple Watch. As the day went on the gap in the step count widened. As of the time of this writing, my Fitbit has 9,994 steps and my Apple Watch is at 7,914. When I look at my Fitbit when walking it counts steps exactly as they happen.

I just purchased a SS after owning a SGS for 9 months. The same issue happened on the SGS. Overall, I''m very happy with my AW, but extremely disappointed in its fitness tracking ability. I guess I'll stick with the Fitbit for fitness tracking and use the AW for everything else. I just don't think Apple is doing a very good job with fitness on both the AW and the Health App.

Am I doing something wrong here? Am I missing anything?

 

[Gav2k]

Have you entered your metrics into the health app and done any of the steps to calibrate the watch because my watch is pretty spot on

 

[dk001]

I use both my Apple Watch and Fitbit One to count steps. What I've noticed is that my Fitbit is much more accurate than my Apple Watch. I've calibrated my Apple Watch using the workout app on my watch and walking with my iPhone for 20 minutes. In fact, I've held the phone in my left hand (I wear my watch on my left wrist) during my entire outdoor walk. I've done this numerous times. What I've found is that the Fitbit does a much better job of counting steps.

Today, I did an experiment where I started wearing my Fitbit at the same time I started wearing my Apple Watch. As the day went on the gap in the step count widened. As of the time of this writing, my Fitbit has 9,994 steps and my Apple Watch is at 7,914. When I look at my Fitbit when walking it counts steps exactly as they happen.

I just purchased a SS after owning a SGS for 9 months. The same issue happened on the SGS. Overall, I''m very happy with my AW, but extremely disappointed in its fitness tracking ability. I guess I'll stick with the Fitbit for fitness tracking and use the AW for everything else. I just don't think Apple is doing a very good job with fitness on both the AW and the Health App.

Am I doing something wrong here? Am I missing anything?

I have found that my FitBit HR is more accurate than anything else in the smartwatch category I have tried.

 

[BSG75]

Have you entered your metrics into the health app and done any of the steps to calibrate the watch because my watch is pretty spot on

Yes, I've entered all of my data in the health app and calibrated the watch following the steps outlined by Apple.
Both the AW and Fitbit are on my left side.

 

[sean000]

I used a Fitbit One for over a year before I bought my Apple Watch in July. I used both for several months so I could see how they compared. Three months of data showed the two devices were reporting very close numbers in terms of step count. On a 10,000 step day the Fitbit and Apple Watch were usually within 300 steps or so of each other.

Sean

 

[Newtons Apple]

Counting steps is not a perfect science. I bet if you wore two Fitbits or two Apple watches, even their counts would not agree. Best to use the step,count as a comparison figure for day to day use.

 

[Julien]

...What I've noticed is that my Fitbit is much more accurate than my Apple Watch....Am I missing anything?

Did you use a manual step counter to confirm that the Fitbit is closer than the Watch to the actual total of steps taken or is it just a subjective opinion?

 

[maflynn]

I can't say which one is more accurate, but from what I've read, all smart watches seem to vary depending on what you use, be it a MS band, fitbit, apple watch etc.

For me, it seem that steps are not promoted as much as the fitbit, which is its primary metric. So I don't really look at my steps now that I've been on the AW.

 

[Julien]

Here is a Cnet comparisons test.

http://www.cnet.com/news/smartwatch-step-counter-and-distance-tracker-accuracy/

 

[Purga]

Detecting a step is not that hard for a fitness tracker. What makes it hard is preventing false positives.
Pedometer, i.e. fitness tracker, are using recurring peaks in accelerometer data to detect steps. This works very well when a person is running or walking, but in everyday life it is harder to distinguish peaks that occur in other activities (lets say brushing your teeth).

true positive:
a person walking which is counting as a step. This is generally very easy to detect from the pocket (iPhone in pocket) and also on the wristband during running. On the wristband (Fitbit / AppleWatch) it is harder to detect when you are doing something else than swinging your arm. For example when you are carrying something.
When you look at the combined data in the Health App of iPhone (pocket) and apple watch (arm), this should be more accurate than the Fitbit which cannot rely on the more stable pocket position. Sometimes you just cannot detect a step from the arm position.

So iPhone + Apple Watch are counting more steps in this regard.

false positive:
Steps are detected when no step occurred. This is due to not optimal algorithms with the step detection.
When testing the Fitbit Flex we realized that a lot of false steps got detected during other activities.

So in this regard the Fitbit is counting more steps.


All in all the counted steps are in a very similar range. However the FitBit is less reliable. It wont detect some steps, but compensate that with detecting too many steps at other occasions.

Just from looking at the amount of steps you can hardly tell which device is more accurate.
It is also not super useful to go for a run and then compare steps. All fitness tracker do quite well in this scenario. It is the use cases that generate false positives and false negatives that are different between trackers.

 

[BSG75]

Did you use a manual step counter to confirm that the Fitbit is closer than the Watch to the actual total of steps taken or is it just a subjective opinion?

Sometimes when I walk I'll hold my Fitbit One in my hand to see how its counting steps. The counter corresponds to each step as it happens. Perhaps there is a difference since the AW is wrist-mounted while the Fitbit One spends most of its time in my pocket. One thing I do like better about the AW fitness app is that is accounts for more than just steps by tracking active calories, exercise and stand hours, while the Fitbit is primarily focused on steps.

It just seemed strange that the AW's step "guess" is so different from the Fitbit's step "guess." Well, as others have said, it's not an exact science.

Thanks for the CNET link.
[doublepost=1461588462][/doublepost]

Detecting a step is not that hard for a fitness tracker. What makes it hard is preventing false positives.
Pedometer, i.e. fitness tracker, are using recurring peaks in accelerometer data to detect steps. This works very well when a person is running or walking, but in everyday life it is harder to distinguish peaks that occur in other activities (lets say brushing your teeth).

true positive:
a person walking which is counting as a step. This is generally very easy to detect from the pocket (iPhone in pocket) and also on the wristband during running. On the wristband (Fitbit / AppleWatch) it is harder to detect when you are doing something else than swinging your arm. For example when you are carrying something.
When you look at the combined data in the Health App of iPhone (pocket) and apple watch (arm), this should be more accurate than the Fitbit which cannot rely on the more stable pocket position. Sometimes you just cannot detect a step from the arm position.

So iPhone + Apple Watch are counting more steps in this regard.

false positive:
Steps are detected when no step occurred. This is due to not optimal algorithms with the step detection.
When testing the Fitbit Flex we realized that a lot of false steps got detected during other activities.

So in this regard the Fitbit is counting more steps.


All in all the counted steps are in a very similar range. However the FitBit is less reliable. It wont detect some steps, but compensate that with detecting too many steps at other occasions.

Just from looking at the amount of steps you can hardly tell which device is more accurate.
It is also not super useful to go for a run and then compare steps. All fitness tracker do quite well in this scenario. It is the use cases that generate false positives and false negatives that are different between trackers.

Thanks, Purga. That's a good explanation of how the devices are tracking steps. I guess I should probably get out of the Fitbit "step mentality" and start focusing on active calories and exercise.

 

[lauramayer]

I used a Fitbit One for over a year before I bought my Apple Watch in July. I used both for several months so I could see how they compared. Three months of data showed the two devices were reporting very close numbers in terms of step count. On a 10,000 step day the Fitbit and Apple Watch were usually within 300 steps or so of each other.

Sean

I can confirm this. My neighbor have only a little difference between both devices.

 

[wolfedude88]

For me personally my fitbit charge and fitbit alta have about 2,000 more steps more than my apple watch. I was shocked to see such a difference between the two.


I haven't calibrated it yet and I just got it a couple of days ago, but does from the way it looks does the apple watch not show you the steps instantly that you get? When I am looking at the screen it doesn't update for a while.

Is that how it is for everyone else?

 

[BSG75]

For me personally my fitbit charge and fitbit alta have about 2,000 more steps more than my apple watch. I was shocked to see such a difference between the two.


I haven't calibrated it yet and I just got it a couple of days ago, but does from the way it looks does the apple watch not show you the steps instantly that you get? When I am looking at the screen it doesn't update for a while.

Is that how it is for everyone else?

That's been the issue for me. Despite multiple calibrations with the AW, the Fitbit always has a higher step count. On walks I will have the Fitbit app open on my iPhone and it is tracking steps exactly as they happen. You are correct that the AW does not count steps in real time, like Fitbit, but shows how many steps you've taken at the time you look at your fitness rings either from a glance on the Watch or open the app on your iPhone.

I think the biggest difference between the AW and Fitbit is the way each device tracks activity. For Fitbit, it's main measurement metric is steps taken. Whether you stroll or walk at a fast pace, it's all the same in the step count, so it turns into a quantity over quality metric.

For the AW, steps are just one metric measured. The AW is more focused on measuring how many active calories you burn, exercise time and stand time in a given day. So from that perspective, the AW gives you a better idea of how active you are throughout your day.

With that being said, I'm using two devices to track my activity. For steps, I'll rely on the Fitbit since this is their area of expertise. For everything else fitness-related, I'm relying on my AW.

 

[Julien]

Keep in mind calibration is for distance and NOT step count. Step count is an absolute but distance is relative to your stride length.

 

[xraydoc]

Also keep in mind the size of the Fitbit and what its (almost only) function is - to count steps - and how much internal volume is dedicated to that.

Now look at the AW and compare its thousand-fold higher general computing requirements doing far more than just counting steps, in what is almost certainly a smaller internal volume.

I'd say if the AW is even close to the Fitbit (and that's presuming the Fitbit is actually accurate), it's doing a good job.

 

[dk001]

Keep in mind calibration is for distance and NOT step count. Step count is an absolute but distance is relative to your stride length.

If I am walking on a flat surface, yes.
I walk in the hills, parks, beaches, tracks, neighborhoods, stairs, etc... My stride length varies. I have found that smartwatches tend to calculate Distance/StrideLength = StepCount whereas the FitBot HR it measures actual steps.

 

[parseckadet]

Counting steps is something Fitbit really latched onto back before heart rate monitoring on these devices became more common. That's when this 10,000 step daily goal notion started, or at least got marketed heavily. But as others have pointed out, not all steps are created equal, and the true goal should be calories. That's why the rings on the Apple Watch set a calorie goal, not a step goal.

Personally, I started discounting the accuracy of every Fitbit when my wife got a Charge HR and it gave her credit for 3,000 steps every day while she was driving to and from work. Kinda easy to hit 10,000 steps when you get 6,000 for free sitting on your butt in traffic. My Apple Watch has never had a problem like that.

 

[mirz2000]

I had a similar experience with the Garmin vivosmart versus the Apple Watch. Here was my end of day results... I notice the Apple Watch misses steps more when I am carrying items in my left hand and am not swinging my arm as much... Whereas the Garmin didn't.

 

[mildocjr]

probably the best way to count steps is to attach a sensor at each ankle, not practical, I get my steps in just reaching for my coffee cup regardless of whether its with a fitbit or AW. Coffee is pretty much the opposite of walking

 

[Purga]

I have found that smartwatches tend to calculate Distance/StrideLength = StepCount whereas the FitBot HR it measures actual steps.

It is actually the other way around. Atleast for the Apple Watch.
Distance = StepCount*StrideLength.
Steps are easier to detect and as a result a binary decision (with potentially false positive and negatives). StrideLength is approximated via the aplitude of the oscillation of the accelerometer data. It is calibrated when running outside (using GPS for distance approximation).
The problem for distance is that it is only measurable via GPS when running outside in more or less straight lines. For all other occasions you need steps and the stride length to approximate the distance
[doublepost=1461709035][/doublepost]

I had a similar experience with the Garmin vivosmart versus the Apple Watch. Here was my end of day results... I notice the Apple Watch misses steps more when I am carrying items in my left hand and am not swinging my arm as much... Whereas the Garmin didn't.

It is better to compare the steps not with the stepcount on the watch, but with the stepcount in the Health-App. The latter combines Watch and Phone data. When you carry stuff the iPhone in your pocket reliably tracks steps (which do not show up on the watch).

Did you do testing with the Garmin and the step count when carrying stuff?
I bet the Garmin doesnt more reliably track steps, but instead has more false positives and due to this a higher count.

 

[dk001]

It is actually the other way around. Atleast for the Apple Watch.
Distance = StepCount*StrideLength.
Steps are easier to detect and as a result a binary decision (with potentially false positive and negatives). StrideLength is approximated via the aplitude of the oscillation of the accelerometer data. It is calibrated when running outside (using GPS for distance approximation).
The problem for distance is that it is only measurable via GPS when running outside in more or less straight lines. For all other occasions you need steps and the stride length to approximate the distance
[doublepost=1461709035][/doublepost]
It is better to compare the steps not with the stepcount on the watch, but with the stepcount in the Health-App. The latter combines Watch and Phone data. When you carry stuff the iPhone in your pocket reliably tracks steps (which do not show up on the watch).

Did you do testing with the Garmin and the step count when carrying stuff?
I bet the Garmin doesnt more reliably track steps, but instead has more false positives and due to this a higher count.

Positive? Both my AW and LG Urbane were far from accurate when not on a flat surface whereas my Fitbit had great accuracy. On a flat surface they all were close.

 

[mirz2000]

It is better to compare the steps not with the stepcount on the watch, but with the stepcount in the Health-App. The latter combines Watch and Phone data. When you carry stuff the iPhone in your pocket reliably tracks steps (which do not show up on the watch).

Did you do testing with the Garmin and the step count when carrying stuff?
I bet the Garmin doesnt more reliably track steps, but instead has more false positives and due to this a higher count.

I don't carry my phone in my pocket because women's clothes don't usually have deep enough pockets to carry much of anything. True story, sometimes we even get completely nonfunctional pockets that just appear to be pockets but are actually just a seam for show. Crazy. But I am sure the phone would be a more accurate reading if one did carry it that way.

 

[MistahMumford]

I typically don't worry much about my step count between my AW and the charge HR that I've not strapped on in days! I'm more concerned with the calorie counts and just being more active

 

[sean000]

I wore my old Fitbit for about 3 months alongside my Apple Watch, and for a 10,000 step day the two were always very close... usually within a few hundred steps. I have a Fitbit One though, which is worn clipped to your clothes or in a pocket, so it's possible that accuracy will vary depending on where you wear it. Fitbit's real-time step counting made me think it was very accurate. I would count the steps as I walked and stop to check and it was always very close. Of course stride length varies, as does effort required, with changes in terrain and steepness. One way the Fitbit got around this was to offer the stairs climbed metric. The iPhone's Health app offers this too, but only when you carry your iPhone. If you climb stairs wearing only the Apple Watch it will track your steps and your heart rate, so at least you might get some exercise and move credit for the effort.

My problem with the Fitbit is that I get a lot of my exercise commuting and running errands by bicycle, so there would be days when I would get a lot of exercise but Fitbit would not accurately reflect that without me starting a workout every time I hopped on my bike. I feel like the Apple Watch gives me a more consistent historical report on my activity and exercise levels whether I biked more or walked more on a particular day. Also, when I did start a workout, my Fitbit was terrible for tracking bicycle rides. The Fitbits that have GPS and HR are probably much better. With the Apple Watch it becomes a combination of the watch's HR sensor and the iPhone's GPS.

So the Fitbit focus on steps wasn't really ideal for me, but it's an easy metric to allow for competition within the Fitbit social network. I feel like the three rings on the Apple Watch are better at reflecting how active I was on a given day.

Sean