iPhone 6 Plus M7 Chip Step Data Accuracy OFF compared to FITBIT One

Downloaded the wristband manager, but oddly enough it will not allow me to sync to Health and keeps giving me a failed sync error and asks me to see if I have authorized Health Kit to allow access. Checking the privacy settings and sources under the Health app seem to show that it is approved so I'm not sure what's going on. Did you have problems getting it to work initially and if so what did you do to get that to work.

No worries, let me know what you find.

Sadly my best guess right now is that it's a software limitation issue. By software I don't mean that it's a glitch, but that the settings and algorithm for using the data from the chip is simply not as smart as the program in the FitBit. People getting good consistency just happen to be holding the phone in a better position or getting most of their data in formats where the FitBit and iPhone are consistent.

It would be nice if there were a way to reset this, but I think a reset really applies to software and not hardware. I have sneaking suspicion that we are both thinking hardware issue. Hopefully I'm wrong and there is actually something in the firmware that can get gummed up and screw up the settings, but somehow I think it's more likely that others who are getting accuracy similar to their Fitbit might just be using the device differently.

Now if it is true this is a software limitation and that the M7 and M8 chips do provide the same data the FitBit gets, then in theory if we can find an iPhone app that actually gets a direct data feed (ie not just copy the step count) and that has a better algorithm to determine what is and is not a step that could be an immediate fix.

I'm still hoping the Apple guys tell you otherwise, that there is something they can reset or give you a new phone. I've had them give me new phones for issues like this before, but they did it more as a one time favor and had no way of testing or verifying that something was really wrong with it (in this particular case it seemed the GPS was not working all the time so it sort of falls into the same area as motion chip).

Karell,

Yes setting the wristband manager to sync all data does work. Sadly then trying to change the preferences afterwards caused the sync to fail again. It would be nice if they mentioned that in the FAQ.

So today's run with the FitBit clipped to the neck of my T shirt and the iPhone in my hand as I like checking to see what my pace is.

FitBit 7,633 steps

iPhone 6,629 steps

What's interesting is prior to this run the iPhone was already about 2,000 steps behind. The approx numbers were FitBit 8,000, iPhone 6,000

That's a pretty significant difference and one could argue that carrying the thing in my hand might have yielded higher numbers, but then it could also have cushioned the steps giving worse numbers. I'll try on my waist next time, but that's not as much fun. In my experience the FitBit is pretty good at registering steps regardless of whether it's in my hand, pocket, or around my neck.

Since I've walked and counted steps using the FitBit and found it to be dead on, I am much more inclined to believe the FitBit number than the iPhone. I've also watched how the FitBit responds to a drive over a bumpy road. It occasionally will register a step, but not very many. It's a good product, probably better than most. I haven't, however, tried to walk around the house and count steps and see how it does, although sometime I watch it as I walk around and it is almost always dead on, whereas the iPhone has this delay before it seems to kick in and sometimes it catches up and sometimes it doesn't seem to.

Tried the reset and really can't tell much of a difference, but its hard to tell unless you actually take some still shots of the data (hold the home button and power button at the same time and wait for the click). I do observe that the darn Health App crashes repeatedly and often doesn't even show the sources or data for long periods of time. Fortunately it appears that the Health App actually draws its data from a separate program that does motion that we don't have access to except in a rudimentary fashion in Settings, Motion. The only reset sadly is to actually reset network settings and then it usually, but not always starts over and the health app works.

I hope that others can report their comparative data this is what I found wearing FitBit on my neck T shirt and the iPhone on a holster on my hip while doing a fairly intense 30' workout on the Elliptical. Keep in mind that neither product is meant to track steps in this scenario and that on my neck with me holding onto the bar the neck position gets the LEAST motion. Fitbit steps determined by setting the iPhone App on track activity and doing a start stop.

Here's the data:

Elliptical ACTUAL STEPS (since it actually meausre this we know its accurate) 4,405

iPhone steps calculated Before 1,672 After 5975 Total 4,303

FitBit (around neck so least accurate position) Total 1,564

My take on this is that the iPhone which was only off by 102 steps is fairly close when worn on the hip and used in a vigorous step action. Prior to getting on the Elliptical, the iPhone was behind the FitBit in step calculation around the house by about 500-1000 steps (which is a lot).

Another Data Point. .....

I think if we all try to record activities and note what we do, where the iPhone and where the FitBit is carried, and how we are determining what is measured by each that will help us better understand what is happening.

Here is the info on a 3 mile walk 20 min per mile mostly flat.

FitBit worn on neck of T Shirt, iPhone worn on holster.

1. iPhone 7,135 steps before 13,039 steps after TOTAL 5,904

2. FitBit 6,740 steps before 12,593 steps after TOTAL 5,853

3. FitBit Calculation from Exercise Monitor TOTAL 6,162

I'm a little surprised that the FitBit feature that monitors your actual exercise actually has a different number than what shows up on the dashboard before and after. I don't have an explanation, but the difference is only about 200 steps out of 6,000 so its not too bad. What is interesting is that the iPhone and FitBit numbers are now within 50 of each other and since the FitBit was around the neck in theory it might have under counted a little.

I think this confirms that the iPhone undercounting problem is NOT likely to be from open walking, running, treadmill etc. It is likely coming from all the steps that are done inside the house, around the office, perhaps walking in a shopping mall or up and down steps, etc etc.

I believe the step counts are made more from the vibration that occurs upon making a step impact. If I wave the FitBit around I don't find it generates steps. While vibration is technically motion, it can be very subtle and its transmission from the feet to the rest of the body is diminished the further from the feet it is. However, this is characteristic and can be detected primarily on the main axis of the body (hence neck isn't too bad, but wrist would be more challenging if not properly calibrated).

Karrell,

Oddly enough your picture didn't come through so if you can resend it or provide the link perhaps we can see what you are looking at.

Here is my data from today, although its a really slow run.

Distance 3.55 miles

FitBit Activity Tracker in App 7,286 steps

FitBit Before 1859 steps After 9,747 steps 7,888 steps (now admittedly I walked to my desk after stopping the tracking, but 600 steps? doubtful)

FitBit Activity Record (press FitBit) 7,785 steps

iPhone Before 2239 steps After 9,885 steps 7,646 steps (I think because wrist band synced the Health App has extra steps from AM)

Withings Before 1422 steps After 9068 steps 7,646 steps

FitBit worn around the neck.

Now because this is a slow run it is hardly faster than a walk (I was tired what can I say).

This data continues to support the idea that iPhone tracking of walking and running when worn around the waist is probably just fine. It also confirms that if you hold the phone in your hand while you are walking that it will under count (ie won't register steps properly). This is undoubtedly why it under counted by 1,000 steps in my previous run. It's only a single data point, but I think its a good bet that's what happened.

The one confounding factor is that I decided this AM to do the following to reset my phone.

1. Backup the phone

2 Delete and erase the phone and set-up as New

3. Just for good measure did a Reset of ALL settings

4. Restored my recent backup (what a pain, but easier than starting over)

It's possible that the Restore and Reset helped get the software/firmware working better, but remember we got good numbers BEFORE this was done. I still think it is much more likely that the problem is with measuring steps around the house where you don't have the same pounding that comes from wearing shoes and are walking on carpet etc. If you are NOT working out and all or most of your steps are indoors my experience is about a 20-30% under count versus the FitBit. But lets collect more information.

Since you cannot test workouts can you test isolated segments of the day where you know where and how you are walking indoors? (ie with or without shoes, wood floors, carpet etc and keep track of where thee FitBit is and where the iPhone is. Try to get at least 2000 steps of data more if you can. All you need to do is make screen shots of the FitBit app and the iPhone Health app step count Before and After. Keep track of whether there are a lot of stairs involved as I think that probably is missed by the iPhone as well. I have no idea why the FitBit app exercise tracking thing measures significantly less steps than the real tracker. I've submitted this observation to FitBit and have not had a reply yet.

Karrell,

I think we are in general agreement that this looks like a software issue. However, in my experience with large companies and without a lot of complaints and/or enough specific information to fix the problem easily they will often say a fix is on the way, but nothing of the sort will happen for months perhaps ever.... Besides if we better understand exactly what the limitations are we can:

1. Help Apple fix this more easily.

2. Perhaps identify an easy work around that enables you to get the data you want.

I know you may not want to include information, but I am hoping others will and that together we might have a better idea of what the problem is and what can be done to help fix it.

Clearly:

1. Holding the iPhone in your hand while walking will result in a significant drop in step counting accuracy.

2. Wearing the iPhone on your waist does work and register well.

However, what about putting it on your arm vs pocket, vs waist? Is the problem primarily with steps or does it occur with start and stops or could it have something to do with wearing or not wearing shoes?

Here is my data on walking around the house barefoot. There were a few times the phone was not on me but that could account for at most about 100 steps.

iPhone Native Tracker 1,860 steps

FitBit Tracker 2,921 steps

Data from the FitBit suggests that I did 12 flights of stairs at this time so that too could have had an impact. Because if you go up 12 and are at the same place you must also have gone down 12 so 24 staircases is a significant portion of the 2,921 steps.

This is about a 1100 step difference or nearly 60%. I think there is little question that the inaccuracy occurs when you are inside the house. Now does barefoot and shoes make a difference?