I’m training for the Copenhagen Marathon in May of this year (2015), and have been using Strava and a new Fitbit Surge device to track my runs. Unfortunately the two systems are not compatible (update May 2015–they are now compatible, details towards end of this post), and I can’t use the heart-rate readings from the Fitbit Surge as health data inside the surge app. I just completed a 15m training run and I thought I would post the side by side data that the two apps gathered in hopes this might be of use to other runners considering using either of these two systems.
Some immediate observations about the differences in the data that was gathered:
- The GPS tracking of the Strava, which is running on my iPhone 6, is much more accurate than the location readings on the Fitbit Surge. The Strava/iPhone readings are really precise and show small variations in my 4 laps around the lake. The Fitbit Surge GPS is almost comically “loose”, suggesting a meandering variation on each lap, sometimes straying into the lake itself or across non-existent streets.
- The splits/pace information is pretty consistent. I tried to start both devices tracking at the same time, but the differences in the splits and the total distance and time of the run may be as a result of slight differences in start time, and, when I paused for a 40 second water break mid-way i manually paused the fitbit clock, but the Strava was on auto-pause and may have taken a different sense of that timeframe. That would account for the 2:05:08 (fitbit) v 2:05:34 (strava).
- No idea how to rationalize the 15.01m Fitbit distance vrs. the 15.3m for Strava, which results in the pace discrepancy: Ftibit says i had a 8:20 overall pace, Strava 8:13 pace.
My sincere desire is that Strava and Fitbit will get their systems connected, not sure who has the burden to do what work, but certainly can’t be very far from core to their missions to support as many devices/APIs as possible in this connected health tracking wearable category? Strava is focussed on community, Fitbit on devices–let’s go guys!
Later update (March 10 2015): I’ve continued to run 4-5 times per week using both Fitbit Surge and Strava and can add some new data:
- The Fitbit Surge is VERY inaccurate in terms of distance travelled on a run, whether that be a 3 mile or a 21 mile course, on road/path or on a treadmill. I’ve seen a consistent 15-20% under-report of distance run in both free-run (GPS tracking on) mode and “treadmill” run mode. In Treadmill mode the distance travelled is under-reported almost comically–the device is just pretty much worthless in accuracy on a treadmill. I would think that the device’s software could compare my GPS enabled and not-enabled runs and correct itself to a better estimate of my gait/pace based on the other data its sensors are gathering. Alas, the software is clearly NOT doing that kind of comparison of the data and optimization to individual user performance.
- The Fitbit Surge is VERY inaccurate in terms of geo-location in GPS mode, and as a result, the pace readings during a run are absurd as well. While running at a 8:15 minutes per mile pace the surge will report anything in a range of 7:45 to 9:30–ostensibly because it has no idea where i am physically on a map. The readings that come back from my runs are hilarious–showing me running through buildings, into lakes, etc.–yes, it is approximately correct, but nowhere near accurate enough to track distance and therefore useless for pace and overall splits.
- While i have no second heart rate monitor tool to compare the readings to, i’m also confident that the hear rate readings are wildly inaccurate, as during a long run where i’m in a steady state zone of pace/energy/effort, the heart rate readings will ramp up and down by 10% range which is attributable to inaccuracy of the device’s readings, not variations in my heart rate load.
Here is a recent run that Strava reported as 21.1 miles, side by side with Fitbit. Notice the accuracy of tracking on Fitbit Surge leaves much to be desired–which makes the device pretty useless as a serious fitness device:
Give my first 2 months of experience with the Fitbit Surge, i would not recommend as a health tracker–way too inaccurate. I’m actually now very curious about the Apple Watch which is shipping soon, given that it will rely on the phone for measurements which i’ve found to be much more accurate (via Strava).
Update May 2015. So a few things happened next for me. On a 15 mile run i tripped and fell face first into a ditch. At the time i had enough natural endorphins pulsing through my body that I just got up and kept running–but within 24hrs i was in incredible pain along my left rib-cage. For next 5 weeks I have been laid up and haven’t been able to run, alas, the travails of training for marathons (this seems to happen to me 50% of the time I train). So i missed my marathons (i had rescheduled my target run to the San Luis Obispo race, since I was ready for a earlier race than the originally planned Coppenhagen). At this point, i’ll be re-starting my training for a late summer date tbd.
But, in that timeframe a lot has happened:
- Fitbit released a patch to firmware that some have asked/suggested might improve accuracy. I have yet to try (will update once I do).
- Fitbit got their act together and there is now compatibility with Strava, http://strava.fitbit.com, which seems to push performance data back and forth between the two systems. Yay for fitbit. I’m going to try soon, have linked my accounts but am not yet pushing real miles through the system so need to get my runs up to 5+ miles for the data to be interesting.
- i got an Apple Watch, and have started using it as yet another ecosystem of data and sensors.
So, here’s my first impressions of the Apple Watch as a fitness and performance sports tracker:
- I miss my Fitbit. Apple’s passive tracking focusses on 3 key performance indicators (KPIs)–how often you stand for 1 minute in an hour of each of 12 hours of a day, how many minutes of “activity” you have in a day, and how many calories you burn in a day. I miss the Fitbit Surge’s focus on steps and distance and stairs/steps, which felt more accurate and meaningful to my daily “activity” goals. The Apple Watch notion of standing for 1 minute of each hour leads to several little notifications throughout the day while i’m at work at the end of a 60 minute meeting that has run over a few minutes… when suddenly not only I, but the other geeks in the room with a new Apple Watch, all get this little buzz on their wrists. I feel like a lemming! The little concentric circles in the UI of the watch, each representing one of the 3 KPIs, is *very* cool visualization, but the KPIs that are being tracked are not for me!
- As a distance/performance tracker, i’ve had limited experience but the Apple Watch application with the green circle and a figure running, which offers tracking for various walk, run, swim, row, pedal type sport workouts, is really pretty lame. It allows distance, calorie, or time targets or Open setting to just track–but it doesn’t track GPS activity on the watch or give splits or any other serious feedback on performance. I think apple will add more integration of this data in the future, or build a companion app for the iPhone… but for now, this is pretty much useless.
- Strava does have a nice handy companion app so that you can start/stop and get other workout data from the watch, while the phone in your pocket running Strava does the real work. This seems likely to be my continued preference, and also gives me a remote on the watch face to control the playback of my iPhone spotify account (which i couldn’t do before without taking the phone out of my fanny pack on long runs). So i thin the Apple Watch is going to be a great REMOTE to control my iPhone, but not a self-contained tracker to replace a wrist worn GPS tracker or activity tracker of any kind. Weird, yet another thing to put on my wrist/pocket while i run which also does NOT actually do all the things I want!
Ok so here we go, 4m run using all three: fitbit surge and Apple Watch on my wrist, Strava running on iOS on my iPhone 6 in my fanny pack:
- Apple Watch exercise app says I ran 3.83m, 32:54 time, 288 calories (active), 61 cal (resting–no idea what that is because I was running the entire time, must be part of run where my heart rate was in a lower range, nay, incorrectly reading as low as I was running fast first mile), 349 total calories, 8:34 pace, average heart rate 175bpm (yikes, i’d be dead if that was true–i’m 44 years old). Pretty poor and inaccurate data.
- Fitbit Surge says i ran 4.00 miles, 32:38 time, 8:09 pace, calories burned 474, and average heard rate 160 bpm and of the 33 minutes heard rate was in “peak” range 30 minutes, cardio range 2 minutes, fat burn 1 minute (seems like much much better heart rate readings than the apple watch.
- Strava says i ran 4.00 miles, 32:40 time, 8:01 pace, burned 686 calories. Not sure how I get the strava/fitbit data to sync, i don’t see the fitbit heartrate data in the strava app… nor did either app post to the other in any other way that I can see. Hmm…
So, 3 pretty messy and different reports. The time differences have to do with me not being able to exactly start and stop them all at once as I had to fiddle with each device. I like the mileage reading from strava+fitbit now showing exactly same figure, which i know from measuring on google-maps to be accurate! that’s an improvement on the fitbit surge with the new firmware (it has never previously reported this run as 4.00m). Here’s the mapping data from fitbit and stava side by side, which looks much better than previous runs (i would still like to test on a longer run and with more varied route, but this does look like it has been fixed by the firmware update last month to fitbit surge!)
In figure above the Strava track is on the left, Fitbit on the right. I can definitely see huge improvements in accuracy and much more frequent samples in the Fitbit Surge data, which actually caught a few subtle route details that were missed by the Strava plot. This is a huge improvement.
Based on this run’s data, if Fitbit and Surge data integration is actually working (still tbd how to make that work) i think i might sell the apple watch to a colleague at work who wants it, and go back to using the combination of fitbit and strava. The apple watch is a turkey–too little battery life (1 day), not enough useful functions, and crappy sports tracker.
I’ll add to this post if I learn more.
Forest Key's Blog 