I have a confession to make: I love my Fitbit.
I’m actually a little bit addicted to it. The little guy does a good job letting me know when I’ve been a sedentary slob – a run-less day sitting at my computer will typically net me only a few thousand steps, which is my indicator to me that I did not move enough. Fitbit does not lie.
I’ve used a Fitbit Ultra for several years now, and I’ve miraculously managed to avoid losing it or destroying it. It has gone missing for days at a time, but always seems to show up again. Most often he reappears in the pocket of a recently washed pair of pants/shorts. His case is now cracking, be he keeps on working. But, I fear that if the cracks continue to grow it might soon be time for the Ultra to be retired.
My only real complaint about the Fitbit Ultra is that it seems to undercount running steps. I’m in Florida right now at Disney World and have been monitoring my steps for a forthcoming post. I’ve determined that I take about 2500 walking steps per mile (my walking cadence is about 110 steps/min). Fitbit seems to do a pretty good job of tallying my walking steps.
Running is a different story. My running cadence averages about 180 steps/min with some variation depending on speed (a bit lower for slower paces, and it can push toward 200 if I’m going sub 6:00/mile pace). If you figure an average pace of about 8:00/mile, then one mile should give me about 1400 steps per mile. As a test, yesterday I ran for exactly 40 minutes (5.19 miles) to Downtown Disney, around Saratoga Springs, and back to Old Key West where I’m staying. The ISmoothRun App recorded an average cadence of 176 steps/min for this run, and over 40 minutes that computes to a bit over 7000 steps (ISmoothRun counted 7149 steps total – held it in my hand on the run). On the same run my Fitbit Ultra clipped to my waist measured only about 4400 steps. So Fitbit appears to be recording only about 60-65% of my running steps.
I’m curious to know if others have found similar running step recording inaccuracy with their Fitbit, and whether any of the newer Fitbit models (Force, One) might perform any better?
I’m also wondering why the recording seems accurate for walking but not for running? And why my iPhone 5S measures running steps fine, but the Fitbit does not (is it just hand vs. waistband placement? – maybe need to test this by running with the Fitbit on my wrist or in hand).
In the grand scheme of things it’s really not a big deal as the real value of the device to me is the reminder that walking more makes a difference to my daily activity total and that a day below 3000 steps is bad, but as a tech junky I do wonder about how these devices work and why they might work well for one type of movement and not another.
Update: Shortly after writing this post I went for a run – 3.27 miles in 26:13. This time I attached the Fitbit to my watch band on my wrist. Result from iPhone in same-side hand was 4659 steps, whereas the Fitbit recorded 4543 steps. Much closer, so seems the loss of Fitbit step-counting accuracy while running has something to do with being placed on the waistband while running.
About Peter Larson
I have the One and it seems pretty accurate, however if you log your run in activities and put in run distance and time, it will give you a better estimate. Additionally, you can go to the Fitbit Dashboard -> Settings -> Personal Info and under “Body Info” enter in your running stride length and walking stride length.
I have the same issue in my ipod nano. I like to heard music when I’m training. But I notice that when I’m walking it counts the steps well, but when I’m running it only count 50-60% of the steps.
I used the S Health app that’s bundled with my Samsung Galaxy phone, and I’ve noticed the same thing. Walking seems about right, but running seems only about 2/3 what I think it should be.
My guess is that the signal coming from the accelerometer is pretty noisy, and it’s not at all obvious which spikes in the signal are actually a step, so the software is biased to towards assuming ~120 steps/minute. I bet if you put it in the dryer or subjected it to random tumbling for a few minutes, it would could a number of “steps” roughly equal to the number of minutes of tumbling times 120.
The Fitbit Flex works really well in terms of steps giving me approximately 2000 steps/mile when walking and around 1800 steps/mile when running, which seems pretty accurate considering my speeds (9:00 – 10:30/mile).
I also know when they brought out the Flex and Force, they did a lot of work updating the sensor algorithms to accommodate standard walking with arms swinging and when carrying or pushing an item (strollers were of specific interest). So, it may just be that a combination of older hardware (since the Ultra is no longer offered) and older algorithms that may be leading to your difference.
Quite possible. Did work much better today on my wrist. Very tempted to get a Force!
Force has been recalled and not on sale for the time being due to the skin reactions users were experiencing. I imagine they will announce the replacement product as soon as they get it figured out.
Our experience is that the accelerometer measurements of step counts are most effective when on the waist and foot, less effective on the wrist/arm/jacket pocket/backpack. Makes sense when you think about it…
I have the Flex; it is very accurate for me. Compares well with my Garmin 610; they end up just a few steps (okay, maybe a dozen or so) off in 3-5 mile runs. I haven’t really checked it during longer runs than that; I assume it remains close. I love the Flex. Had a Zip, but it perished in the wash.
Travis,
It still compares fairly well. I wore mine for my marathon in November, and a number of longer training runs before that. On marathon day, I logged 52,859 steps over 29.83 total miles for the day according to Fitbit. That gave me an average 1772 steps/mile, which is pretty close to the 1800 steps/mile I normally see on my shorter runs.
Though, I’m far more interested in the distance being right than the number of steps, so I have my Flex lights set to be 1 light per mile.
I have always had the same under-counted step issue with my Fitbit Zip. After adding the FitDataSync app (http://www.fitbit.com/apps/fitdatasync), I noticed that my steps were automatically increased or re-calibrated after syncing with my Garmin Connect account. It doesn’t solve the underlying issue that you’ve raised, but it makes me feel like I’m getting a more accurate reading.
I upgraded from this model to the flex about a year ago and I have found that while force is also not 100% accurate it does track running much better. On shorter runs is a little off. So for example if I run 4 miles, it may only show 3.75. However, the longer the run, the more it’s off. For example, my 12+ mile run this morning only showed up as 10.8 or so. For me it’s not a big deal because I log most execerises manually (it doesn’t track cycling or swimming) and only use fitbit for my daily activities outside of that.
I found exactly the same issue with my Ultra (torso mounted). I reported in my review that it was measuring about 60% of the distance and a cadence of 120-140 spm. (My FR620 reports 180-190). I’ve since upgraded to a One, but I haven’t done any more detailed tests. I hadn’t tried wrist-mount, since those models are (presumably) calibrated for torso-mount.
http://mikesk8s.wordpress.com/2011/11/27/fitbit-ultra-review/
I had a Fitbit until the Kinvara 4 came out so I could not resist selling it on eBay to get the K4. I should have kept the Fitbit. K4 was a bad update.
Peter, I was wondering what case / strap you use for your iPhone when running. Just recently got my very first smartphone, and must say I’m quite paranoid about it falling out of whatever I carry it in (while running). :)
I prefer to hand hold mine. Typically use just a Go Strap affixed to the back of my iPhone case. It’s an elasticized band the I can slide my finger through while I grip the phone. HB Tune also makes good hand held cases.
Cheers, I’ll check them out.
Thor,
I really like the Armpockets (http://www.armpocket.com/). I’ve tried several different options, but none has been as useful for carrying other objects (keys, ID, credit cards, etc.) or as comfortable as the Armpocket.
Thor,
I just got a FlipBelt for my iPhone – comes in black and colors – wear it around your hips and your phone plus ID or key can also go in. I like it very much. Easy access and no movement.
I just got a FlipBelt for my iPhone – comes in black and colors – wear it around your hips and your phone plus ID or key can also go in. I like it very much. Easy access and no movement.
My Fitbit Ultra seems to be dead on. I’ve used it for two 50 mile ultras and it’s right around 100,000 steps for the day (but added about 2 miles to the total). A marathon was 46,000 steps and about 0.5 miles short.
I still prefer the clip-type fitbits versus the watches for accuracy.
Pete, thanks for the update on using the FitBit on your wrist.
I use the FitBit Zip and have noticed it’s most accurate when clipped into my front shorts pocket, about 10 inches down from my waist, but will try clipping to my watch and see how it compares.
No problem Steve! Unfortunately my Fitbit has gone missing since writing this post. It seems to enjoy disappearing from time to time….
I have a similar problem with my Fitbit zip. I clip it to the inside of my running shorts and it always seems to come up a bit short on distance. The discrepancy is particularly noticeable on a treadmill, where a 5 mile run might record about 4.4 miles on my Fitbit, but I have the same issues running on the street. Strangely enough, when I jog alongside my girlfriend, who is the same height as me, and we go exactly the same distance, my Fitbit will record about 90 to 95 percent the distance recorded by my girlfriend’s Fitbit. I’m beginning to get the impression that this little $60 trinket is not a precision instrument. (I’ve also had some issues with the battery.) Should I have paid more to get a better model or should I have bought from a different company? Did I get a lemon, or am I just expecting too much out my Zip? Any thoughts or suggestions would be appreciated.
I think the reality is none of these devices are going to be 100% accurate, and how accurate they are may vary from person to person based on gait, body size, etc. I’m now comparing the Garmin Vivofit and the Garmin FR15 and two devices from the same company are even giving me different results. I think the best way to view these devices is as a general indicator of how active you have been on a given day, not necessarily as an indicator of exactly how many steps you have taken.
I have been using my zip for about a year, and recently decided to use it as a tool to track my progress in in erasing my pace. The funny thing is that as my actual pace went from about 150 steps per minute to 180, my zip records a drop from 150 to about 130 steps per minute. So I attribute part of the inaccuracy to the smoother motion that comes from not over striding.
I keep mine in the pocket at my waist while jogging. Will try different locations next. I also compared my zip with my wife’s fitbit One. Put both of them in my waist pocket. Got essentially the same wrong results.
I’ve been using a Garmin Vivofit for several months now and it has been much more accurate at tracking running steps.
Thanks, Peter. When my FitBit dies, I will look into it. I am so tied into the FitBit universe, that I don’t know if I could switch!
After reading the comments here, I tried running with my Zip in different locations. On one day, I held my Zip in my right hand and my old iPod nano in my left hand. While better, the Zip still undercounted my steps. The iPod appears to have counted only half my steps.
I like to run with a Halo headband, which is tight enough that I can tuck my Zip inside. Running with the Zip on my head resulted in a step count that looks reasonable. It was the first time I have ever seen my Zip record enough steps to show that I had a pace of 180 steps per minute.
I wear my Zip on my waistband while running and it is always under the distance measured by my Garmin which I know is accurate. I ran 8 miles over the weekend and the Fitbit only counted steps for 6.9 miles. I just don’t think it can turn over fast enough. I’ll try it in another location. It’s great for walking and other activities, though.
Karoline-
The fitbit, and all pedometers, calculate the distance based on the measured steps and your stride length. Since you run with a Garmin, you can use that distance to calculate your stride length (distance divided by steps, and convert the result into feet and inches). Plug this number into the settings in Fitbit’s web page.
Of course, it may also be missing steps, which means that your calculation for stride length would be too big.
The other issue will be that if you speed up or slow down, or run on different terrain than when you calibrated, the measures will be off as well.
I have the same problem! I have the One and it is amazing…except when it comes to running. For my 8-9 mile runs, I’ve noticed it only shows up as ~7 miles; sometimes more, sometimes less. I’ve tried attaching it to my waistband and also attaching it to the top of my sports bra (which I realize you probably can’t do
maybe I’ll try to buy one
I have the Surge, and have the same issue. The people at FitBit told me to track my strides manually and that would be the way to be most accurate. My experience with both Garmin and Polar were that they were consistently more accurate, but obviously without all of the fun features.