Step counter is way off

KahneFan

Well-known member
Jul 17, 2011
536
16
0
I have always found my FitBit One to be very accurate. So, I decided to wear my FB1 and my S2 3G yesterday to test the S2. End of the day, S2 5000 steps, FB1 7200 steps. I would imagine since I hold clipboards/papers/babies/groceries in the arm with my watch, it may not be counting steps since my arm is stable as I'm walking? I clip my FB1 to my pocket so it is able to count steps no matter what my arms are doing. Kind of a bummer since I was somewhat rationalizing the cost of this watch by thinking I could combine devices.
 
Some have said that the S2 does not start to count steps until you do at least10. So if you do a bunch of small steps, it may not count them. From what I've seen from my fitbit it counts all of them.

Posted via the Android Central App
 
I have always found my FitBit One to be very accurate. So, I decided to wear my FB1 and my S2 3G yesterday to test the S2. End of the day, S2 5000 steps, FB1 7200 steps. I would imagine since I hold clipboards/papers/babies/groceries in the arm with my watch, it may not be counting steps since my arm is stable as I'm walking? I clip my FB1 to my pocket so it is able to count steps no matter what my arms are doing. Kind of a bummer since I was somewhat rationalizing the cost of this watch by thinking I could combine devices.

Using my fitbit HR as well and noticed that my S2 step count is always reads higher than my fitbit.
Now does that mean my S2 is inaccurate? Not necessarily, because I'm pretty sure that both use different algorithms for their count, not to forget that these products are not the same.
So to say that fitbit is accurate and the S2 isn't, is bit of a stretch without having something that could accurately measure both devices.
I don't understand what you meant by "combining " the 2 devices.

Posted via the Android Central App
 
I don't understand what you meant by "combining " the 2 devices.

Posted via the Android Central App

I had a Pebble + Fitbit, hoping the S2 would combine those 2 and I would only need 1 device. Before the test I was already OK with a 10% difference, but 2200 steps is a really big difference. I have a Charge (not HR) as well. Maybe I will wear both on the same arm tomorrow and compare those two readings. My Charge and my One both sync with my phone, so I don't know that I could swing all 3 since my One and Charge would sync together during the day, unless I simply uninstall Fitbit from my phone for the day.
 
I have a fitbit flex and my fitbit is always approx 2k steps more than the S2. I have actually seen where I was walking and my S2 was not counting steps, and it was way more than 10 steps in the walk smh
 
Anyone comparing their S2 to some other fitness tracker (Fitbit or whatever) is not presenting any real basis for declaring the S2 inaccurate.

Most mainstream trackers are 15-40% off. Fitbit included.
Fitness Trackers Up To 40% Inaccurate: Fitbit, Jawbone, Nike, Others Tested in New Study, Which Performed Best? : PERSONAL TECH : Tech Times

If you want to declare the S2 inaccurate, walk and count off 500 steps and then see where the S2 stands. With the available tech in these devices, it's not a question of if they are inaccurate, it's a question of how much they are off.
 
Anyone comparing their S2 to some other fitness tracker (Fitbit or whatever) is not presenting any real basis for declaring the S2 inaccurate.

Most mainstream trackers are 15-40% off. Fitbit included.
Fitness Trackers Up To 40% Inaccurate: Fitbit, Jawbone, Nike, Others Tested in New Study, Which Performed Best? : PERSONAL TECH : Tech Times

If you want to declare the S2 inaccurate, walk and count off 500 steps and then see where the S2 stands. With the available tech in these devices, it's not a question of if they are inaccurate, it's a question of how much they are off.

nice article, but reading with my 3rd eye, I found flaw with this article on a couple of things.

1) In this thread we're specifically talking about "Step Counting", which this article doesn't talk about and the 2nd paragraph was very vague:

"Researcher's results showed all the trackers to be between 15 to 40 percent off in tracking user activity."

Which wasn't defined until:

Iowa State University did a study to determine exactly how accurate 7 popular fitness trackers were at correctly assessing the calories burned by those wearing them.

2) Article is almost a year old

But I do agree with your 500 step test to guage on which device is accurately counting your steps.

I know my S2 it gives me a higher step count and I wear that on my left wrist and my fitbit on my right. Now whether or not this effects my step count is another question.
 
I had a Pebble + Fitbit, hoping the S2 would combine those 2 and I would only need 1 device. Before the test I was already OK with a 10% difference, but 2200 steps is a really big difference. I have a Charge (not HR) as well. Maybe I will wear both on the same arm tomorrow and compare those two readings. My Charge and my One both sync with my phone, so I don't know that I could swing all 3 since my One and Charge would sync together during the day, unless I simply uninstall Fitbit from my phone for the day.

Samsung doesn't play well with 3rd party hardware, as well as fitbit doesn't play well with others. Like the fitbit app and still using my S Health app, with no conflicts.
 
Samsung doesn't play well with 3rd party hardware, as well as fitbit doesn't play well with others. Like the fitbit app and still using my S Health app, with no conflicts.

I'm not using a 3rd part app with the Samsung. I was speaking of my Fitbit One and my Fitbit Charge syncing together via the Fitbit app on my phone (which I have removed for the day). My Sammy is all alone using it's native apps. For that matter, the only "apps" I've installed on my S2 in the 3 weeks I've had it have been watch faces, so it's pretty much factory based as it sits.
 
...walk and count off 500 steps and then see where the S2 stands...

OK, I did a step test. I will say, the first 350 steps, I feel very accurate on my count. The second set of steps, I saw something shiny and got distracted so I may be off by about 5 steps myself, but it's pretty close.

Device: [Starting count], [350 steps], [370 steps] (with my arm stationary)

S2 (3G): 281, 629 (348), 995 (366)

FB Charge (not HR): 305, 654 (349), 1028 (374)

FB One: 373, 723 (350), 1092 (369)

Doing the step count, they are pretty darn close. However, I will add, I put all 3 devices on this morning at the exact same time and yet they all had different starting points for this test - and not small difference either considering how low the steps were. So, while the step count is good, there is something in (my) real world practice which is throwing these off. I wear my S2 and my Charge on my arm, but I wear my One on my pants pockets. Seeing as the One is higher in step count, it could be counting steps while I've driving. But, I tested that previously and after a 20 mile trip it only registered about 5 steps. So, I don't think my 5 mile drive this morning would make my One run that much higher.
 
I have a previous shattered heel, and monitoring steps is an hourly obsession for me. I literally keep a running count of them in my head for hours upon end, so it is pretty easy for me to compare my actual steps to wearables. My Gear Live (android wear) is consistently closest to my actual steps. In July, I kept a 30 day journal, and it was about 3% under to actual. My S2 undercounts by about 20%. My Gear Fit undercounts by about 15%. My LG Urbane 2nd edition LTE undercounts by about 15%.
 
I wore both a Microsoft Band 2 and my Gear S2 3G for an hour and both devices registered about 500 steps.
 
I previously had a Withings Pulse for which I checked the accuracy by physically counting my steps on multiple occasions. It was virtually spot on all the time. When I switched to my Sammy S2 I used both for a while and found the Sammy consistently under-counts my steps by about 10% regardless of how far I walk. I don't think it is about a difference in algorithms. The Sammy S2 freezes from time to time. When it eventually gets back to work the steps taken during the freeze are lost. I've come to accept this because I love the Sammy but also because I am hoping the software will get a fix.