As I said in my previous blog, my new Fitbit Zip is proving to be an amazing motivator when it comes to exercising. I now find myself going out of my way to walk a few more steps, just to see them reflected on my personal online dashboard. Having said this, there are a few considerations of which users should be aware...
First let us remind ourselves that there are three members of the Fitbit Family, as follows:
* Fitbit Flex: Wireless Activity + Sleep Wristband
* Fitbit One: Wireless Activity + Sleep Tracker
* Fitbit Zip: Wireless Activity Tracker
Each is worn and used in a different way. I have the Fitbit Zip as shown below. This comes with a clip (not shown in this image) that allows you to attach your Fitbit just about anywhere on your person.
As an aside, I would like to note that the Fitbit's clip is a wonder of modern materials technology and engineering. It really grips your clothing, making it practically impossible for you to lose your Fitbit, even if you are running at high speed.
Now, before we proceed to the main thrust of this blog, I must say that I was a little surprised to get the following feedback:
You were walking on the treadmill while writing your post and the Fitbit on your wrist. So your arm was not moving/swinging with every step you were taking? And you expect the Fitbit to record your step without your arm swinging? What nonsense is that? The Fitbit and every other similar arm/wrist band pedometer that pretends to count your step while sitting on your arm is making fools of consumers like you. And once the step count is wrong, every other information such an instrument gives out is wrong. Did this never occur to you?
Well, someone certainly got out of the wrong site of the bed this morning, didn't they? Having said this, after re-reading the above comment a few times, I'm beginning to wonder if the poster's antagonistic and resentful tone is based on the fact that he or she spent some time walking on a treadmill while wearing an armband activity recorder without swinging his or her arm, and that he or she is still disgruntled by the fact that those steps weren't counted towards that day's total (LOL).
I replied to the poster saying: "Did I say the Fitbit was on my wrist? (No) Do I look like an idiot? (Don't answer that :-)"
The thing is that "I R an engineer," so the potential for these problems had—of course—occurred to me. This is why I had already investigated the situation as described below.
The first thing that struck me when I donned my Fitbit Zip was that you wear this little beauty throughout the day and there's no way to turn it on and off. The reason this was of interest to me is that my ride home involves a bumpy road that seems to breed potholes, so I wondered if driving in my truck would trick my Fitbit into thinking I was walking or running.
All I can say is that there is some serious DSP (digital signal processing) taking place. I checked the step count before I set off and again when I arrived home. After travelling more than 10 miles over a very boisterous route in the "bumps department," my Fitbit Zip had only incremented its count by two steps. Personally, I think this is little short of miraculous.
The next morning when I came into work, I decided to try a few things. My typical garb in the summer is shorts, sandals, and a Hawaiian shirt. My leather sandals tend to flap around a bit, much like flip flops. Keeping this picture in mind, I noted the current step count and then started to walk around my office building with the Fitbit clipped to my shirt's breast pocket. I counted 100 steps and then stopped and checked the count against my Fitbit. Sad to relate, there was a noticeable error.
I then repeated the experiment with the Fitbit attached to different parts of my body (well, to different parts of my clothing) as illustrated below:
I might note that all this took a lot longer than you might expect. I didn't have a piece of paper or a pen with me, so I was doing it all in my head (which is why I don't recall the exact values). Also, I'm easily distracted, so I'd see something interesting and lose my count, or I'd stop at 100 steps and realise I couldn't recall the starting value. I'm sure the other folks in the building (a) got tired of seeing me walking past their windows and (b) wondered what the heck I was doing.
The bottom line is that in each case (Fitbit clipped to shirt pocket, short centre, trousers side picket, and trousers centre), there was some amount of error in the step count. I found this to be a little disquieting. I'm a bit of an obsessive compulsive when it comes to things like this. You can call me old-fashioned if you will, but I do like my numbers to add up to the correct value.
Of course, most folks don't perform serious training in flip flops, so my next step (no pun intended) was to exchange my sandals for socks and sneakers and to repeat the experiment. In this case, I'm delighted to say that my Fitbit Zip was pretty much 100 per cent accurate irrespective of the location to which it was attached—I don't think it ever miscounted by more than one step per test.
Now, this is where things get interesting, because I then repeated the entire suite of tests while walking on the desk treadmill in my office. What do you suppose I saw?
I have to admit that I wonder if Truthfinder is reading this column. I still hear (in a metaphorical sense, of course) his or her words ringing in my ears: "Did this never occur to you?" I must admit to feeling a tad miffed by this; perhaps even a mite peeved, if you will pardon my language. (See also England raises threat level from "Miffed" to "Peeved")
So, just to remind ourselves, the following tests were performed on the desk treadmill in my office. This treadmill is very accurate when it comes to counting one's steps. The great thing is that I can watch it counting as I walk, so I have 100 per cent confidence in the values reported by the treadmill.
Once again, as reflected in the image below, I performed the tests with the Fitbit clipped to four locations: my shirt pocket, the middle of my shirt, my trousers side pocket, and the middle of my trousers. The reason for the different amounts of time (number of steps on the treadmill) for each test is that I had to keep on stopping to answer the phone.
I calculated the delta percentage error in the step count using the absolute value of the treadmill count (which we know to be correct) minus the Fitbit's count, all divided by the Fitbit's count and then multiplied by 100.
Now, remember that all four Fitbit-attachment locations reported accurate values when I was walking around the car park on a solid (asphalt) surface. In some cases, however, walking on the treadmill causes the Fitbit to "throw a wobbly" as it were. The worst-case scenario occurs when the Fitbit is clipped to one's shirt pocket (a whopping 55.57 per cent error). The best-case scenario is when the Fitbit is clipped to the centre of one's trousers (a miniscule 0.1 per cent error).
Personally, I am more than happy with a 2-step (0.1 per cent) error out of 1,936 steps. I can live with that. I should, however, note one last point, which is that all of my tests were performed while walking—it may well be that the picture changes when one is jogging or running. Consider the first test shown above in which the Fitbit was attached to my shirt pocket. While walking on the treadmill, the Fitbit reported far fewer steps than I actually took. It may be that, if one were jogging—which can cause a Fitbit attached to one's shirt pocket to bounce up and down a few times—the Fitbit might end up reporting substantially more steps than one had taken. But since I am not a big fan of jogging (you have to look after your knees), I'll leave that experiment to someone else. If you decide to perform that experiment yourself, please share the results with the rest of us.
Well, I think that will do for now. Please comment below if you have any questions or thoughts (preferably on the topic of wearable electronics, but I'm happy to talk about almost any topic—the real trick is to get me to stop talking). Until next time, have a good one!