Underestimating mileage

[Adele287]

I have been using the Fitflex since October 2013 and when compared to the Body Media tracker and actual "driven mileage" walking outside, it was very accurate. Since the software upgrade several months ago, it has been underestimating steps, miles and active minutes! I thought it was because my Fitflex was older; however, just purchased one for my husband and his is inaccurate as well. I'm the Director of a weight loss medical practice and have required patients to purchase this product and they are complaining about the same issue. Bottom Line, I purchased a Jawbone Up for comparison and it was 100% accurate with mileage and active minutes. And, I had performed a sync and reset on both Fitflex trackers before we walked, so that variable is off the table. All fitbit customer service says is "we have referred this problem to our engineers and will get back to you". I have every color band and preordered the Tory Burch one and feel responsible to my patients for recommending the product. Is the premium program more accurate? Very frustrated as I luved my Fitflex, but can't recommend an inaccurate product. Has anyone tried the premium upgrade? Thanks!

[ChandlerBob]

Many users keep talking about "stride," and not steps! I have several devices besides the Alta (Garmin, FB One, Jawbone, etc., - along with spouse's). Fitbit Alta is always about 20% on the low side of STEPS. In other words, it doesn't accurately count STEPS, according to a comparison among other pedometers we own.

I believe there should  be an opion in our "preferences" to calibrate our individual devices. Once the STEPS are correct -- or close to correct -- then we would measure "stride."  Make sense?  

To verify this error, all I do is put three devices on my wrist and walk while manually counting the steps, and then compare all devices. Alta seems to be continually low, by aboiut 20% low on average.

This afternoon, I've been wearing the Fitbit One and the Alta for a couple of hours while working. The count on the One is 2381 steps, while the Alta registers 1613 steps -- in the same amoount of time.

I'm a "step counter," because I know how much distance traveled on different devices, based on steps,, i.e., treadmill, orbiter, and walking.  I'd also like a less busy "dashboard" too, but that's for another post 🙂

[USAF-Larry]

The number of steps and the stride length are used to calculate the distance, @ChandlerBob.  The steps are counted by the arm swing motion - each swing of the arm results in 2 steps.  The total number of steps are then multiplied by the stride length to calculate the distance.

If the step count is not accurate, it is most likely because the arm is not freely swinging, i.e., carrying a water bottle or walking stick will restrict arm motion and result in steps not being counted. I don't understand why different trackers on the same arm would show different counts of steps.

[ChandlerBob]

Hi Larry,

"I don't understand why different trackers on the same arm would show different counts of steps."

That makes two of us 😉
Enjoy your evening!

Bob

Bob Hanshaw - USN (Ret)

Bob Hanshaw
www.bobhanshaw.com
www.linkedin.com/in/bobhanshaw/en

[Heybales]

---

@USAF-Larry wrote:

The number of steps and the stride length are used to calculate the distance, @ChandlerBob.  The steps are counted by the arm swing motion - each swing of the arm results in 2 steps.  The total number of steps are then multiplied by the stride length to calculate the distance.

If the step count is not accurate, it is most likely because the arm is not freely swinging, i.e., carrying a water bottle or walking stick will restrict arm motion and result in steps not being counted. I don't understand why different trackers on the same arm would show different counts of steps.

---

Actually - it's not the swing of the arm.

It's the impact of the steps.

And Fitbit and others is trying to discern that impact and amount and hang time - despite it swinging on an arm.

Which means you grip a shopping cart a certain way and no steps are seen.

Or you cross your arms to your chest and steps will still be seen.

The amount of impact compared to what was expected for the weight and stride length setting gives a dynamic actually used stride length for that step. Those distances add up and calories burn is based on rather accurate formula's for weight and pace and basic metabolism (BMR).

That's why stride length walking should not really be set for your fast exercise pace - which is used for how many hours in the day compared to your normal walking pace?

Neither should you use the grocery store shuffle pace - but somewhere inbetween.

Fitbit then will dynamically adjust both directions - obviously losing accuracy on the outside of the range.

It's also why arm movement can fool it and add steps when not moving - but if you look at the distance given and calories assigned for such low impacts - it usually balances out with missed steps.

[Heybales]

---

@ChandlerBob wrote:
Hi Larry,

"I don't understand why different trackers on the same arm would show different counts of steps."

That makes two of us 😉
Enjoy your evening!

Bob

Bob Hanshaw - USN (Ret)

Bob Hanshaw
www.bobhanshaw.com
www.linkedin.com/in/bobhanshaw/en

---

If by the same vendor - shouldn't be too different - though they do change acceloremeter chipsets sometimes - and those have different accuracy.

Different vendors - the formulas are proprietary to discern foot impact despite the arm swings - or some cheaper ones mays go for merely arm swings.

[USAF-Larry]

---

@Heybales wrote:

Actually - it's not the swing of the arm.

It's the impact of the steps.

---

This is not correct.  You can get steps by simply swinging your arm, or brushing your teeth, without your feet moving or impacting anything.  You may also be able to get some steps by stomping your feet without moving your arms, but stomping your feet will cause your body to twist and some - not all, but some - steps will be counted.  If you would like to test this, simply sync your phone app so that it shows "connected" and then start swinging your arm with the tracker and watch the steps start counting up.

When I am doing curls with a dumbbell, the tracker will count steps when curling the arm with the tracker, and it will not count steps when doing the other arm.

I learned when I was walking with a cane that I could not wear the Charge HR tracker on the arm that I used the cane with, because the arm did not swing enough to count all of the steps. When I moved the tracker to the other arm, it accurately counted the steps. My feet were experiencing exactly the same "impact" regardless of which arm the tracker was on, but it did not accurately count steps when it was on the arm that was using the cane for balance.

[Heybales]

You didn't read my last sentence.

Yes, it'll count "steps" of minimal impact doing non-step stuff.

You can also swing your arms exaggerated and high in such a way as to minimize the impact of your steps and make it think you are taking small light steps - exactly the opposite as you'd expect if the swing was the basis for the step.

Do some tests. How would the swing of the arm cause a change in distance seen?

Walk with your arms crossed to your chest as I mentioned, normal as possible steps (normal for not swinging arms anyway) - you'll still get steps.

You should have gotten one step with the cane arm - but it should have failed to see the other foot coming down since it was solidly on the ground resting on the cane.

I just tested it with my handmade cane used during ankle recovery.

This is the same effect for people that put it on an ankle for walking - since one foot while walking is always on the ground - it'll only see the impact of one foot hiting - unless you are a toe walker bouncing with each step - in which case the distance is usually badly off.

[anderpam]

I am very disappointed in the mileage estimation of the fit bit Blaze. I have only had it for 8 months and it has been getting consistently worse. It used to track my mileage rather closely but is now off by over 25%. I get very frustrated and I have had no luck with the tech support. I have calibrated my steps, have tracked with GPS, and have done everything they have recommended and had no luck. I have always had such good luck with Fit Bit but I feel like I have not had any support lately. Why own a tracker if it does not do it's job!  I researched this purchase extensively and chose this tracker for it's features that I use regularly (treadmill especially!). The tech support told me that it doesn't really track well on a treadmill... why would it have a feature for the treadmill then? And I have done everything I can to make it fit to my stride length, etc... I am ready to throw it away and get something else.

[Heybales]

---

@anderpam wrote:

I am very disappointed in the mileage estimation of the fit bit Blaze. I have only had it for 8 months and it has been getting consistently worse. It used to track my mileage rather closely but is now off by over 25%. I get very frustrated and I have had no luck with the tech support. I have calibrated my steps, have tracked with GPS, and have done everything they have recommended and had no luck. I have always had such good luck with Fit Bit but I feel like I have not had any support lately. Why own a tracker if it does not do it's job!  I researched this purchase extensively and chose this tracker for it's features that I use regularly (treadmill especially!). The tech support told me that it doesn't really track well on a treadmill... why would it have a feature for the treadmill then? And I have done everything I can to make it fit to my stride length, etc... I am ready to throw it away and get something else.

---

So for treadmill the GPS is meaningless.

It doesn't have a "feature" for treadmill - that is merely a text label you can select compared to some other exercise name.

It will still be attempting to see impacts to estimate distance.

So from that, have you recently walked say 100 right foot steps, and seen if it increased by the 200 it should have?

Could be some accuracy of seeing impacts is worse.

Since treadmills must be calibrated for accuracy and I've rarely seen that happen in a gym - have you gone out to say a known distance track with GPS off, and seen how accurate it is?

Since the walking stride length serves better accuracy for the entirety of the day if set to avg daily pace (not grocery store shuffle, not exercise level pace), have you set the distance for that avg - or attempting to nail exercise level pace distance?

If lost weight, could be walking differently too. Have you retest that stride length with about 1/4 mile (400 m) track to confirm at that avg daily pace.

I'll admit, it's hard to just go say 3mph when on track when it feels like should be striding out serious.