There’s a double edged sword going on with modern smartphones. On the one hand they can do loads of stuff, but the flip side is that they might not do certain tasks as well as a dedicated device. For most tasks the compromise might be acceptable, but what are they like for tracking your bike rides? How do they measure up to dedicated GPS units?
In this post I’ll share the findings of a little experiment I did, which involved my friend’s iPhone 4S running the Strava app, my Android phone, a Samsung Galaxy S2, running the MyTracks app, and my Garmin Edge 800.
The experiment was initially devised to determine if Strava segments were being interpreted differently according to variation in the gpx tracks produced by different devices. The idea was for a group of riders to ascend a hill together, and see if strava interpreted the segment timings correctly.
This post is an off shoot of that experiment, mainly because I thought the issue was worth highlighting.
Here is the Garmin track, which we can use as a reference. It follows the road perfectly, and all segments are picked up by Strava, every time.
Here’s the iPhone 4S track. It’s all over the place. It barely follows the road any point, and Strava did not recognise this as the segment it actually covered.
If this is the quality of the GPS tracking on iPhones – don’t use one!
Just stick to Angry Birds or whatever….
Here’s the track recorded by the MyTracks app on my Samsung Galaxy S2 (Android).
It’s very accurate, and like the Garmin track, never deviates from the actual route taken. Strava recognised the segment and ranked it as an identical time to the Garmin.
(The garmin was mounted to my bar, and the phone in my jersey pocket)
I took a closer look at that iPhone track, and found this little delight (below).
I have no idea what’s going on here. We were travelling from the East, following the road by the side of the river. The iPhone starts daydreaming and wanders off the track, only to snap out of it and suddenly jump back to the road at a later point (the zig zag bit).
The garmin and android tracks had no such oddities, and were consistent for the entire 58 mile ride.
If you’re serious about tracking your cycling, there is no better option than a dedicated unit such as a Garmin. Smartphones can give you a taste of it, and on the basis of this limited evidence, it would appear that Android does it more effectively than Apple. But they would both suffer poor battery life anyway, so the Garmin still wins.
There are loads more reasons to ditch your phone and get a proper GPS for your bike.
Hi Alan
Interesting findings. Hope you don’t mind if I put a some what different slant on your comments.
Today I road the Liverpool-Chester-Liverpool 50 mile event and recorder the ride on my Garmin 800. I then downloaded the results to my PC using both Garmin Connect and Strava. I rather surprised to see different results from both.
Garmin Connect Strava
Distance 51.00 miles 51.00 miles
Time 3:15:11 2:59:45
Moving Time 3:14:51
Elapsed Time 3:40:44 3:40:48
Max Speed 27.9 m/h 27.7m/h
Elevation Gain1608 ft 1318 ft
Avg HR 128 129
Max HR 161 161
Sorry for that Alan, I do not know what I did wrong there but I seemed to have posted it without finishing my comment. Possibly when I tried to put the data into columns. The second figures are from Strava.
The Garmin Connect gave me an average speed of 15.7 m/h against Strava at 17.0 m/h. I like Strava’s figure better! There is considerable difference between some of the data yet some is more or less the same. Makes me wonder what is most accurate.
Like your Website.
Hi Ian, thanks for your kind words and for sharing your data. It does seem like several people are reporting differences between Garmin connect and Strava too. Perhaps until we get the next generation of GPS which civilians can access, we’ll always have these discrepancies. Until then, I’ll keep telling myself it’s reliable *enough* because it motivates me to train harder.
Alan
Re: altitude – I would suggest enabling the elevation correction function in Garmin connect and see if that brings the figures closer together. The Edge 800 calculates elevation on the fly using barometric pressure, and if it’s anything like my Edge 500, it’ll produce some fairly odd looking results in wet/foggy/humid conditions – especially if you pop into the clouds.
Re: average speeds – my guess is that Strava has calculated on the basis of moving time whereas Garmin Connect has calculated in time. I assume there’s some number crunching going on to determine whether you’re moving or not – presumably a setting somewhere that says if you’re travlling at less than x, then the data is erroneous. I often see my Edge 500 floating around 0kph and 1kph when I’m stopped at a junction, for example.
Hi Alan,
Interesting notes. I recently did the Wiggle French Revolution ride in around northern France. I used my new Garmin 800, one friend used his Garmin 500 and other friends used the Strava app on their various smartphones.
When we all uploaded our times mine registered that I’d climbed 800ft more than everyone else, which I had not. There was a difference of about 50 / 80ft between Garmin 500 and all the smartphones (none were exactly the same), but I am puzzled as to why mine was inaccurate. The distance was the same. Any suggestions on how to collaborate it?
Thomas
I rode with a guy that had a Garmin 550. My Strava app failed during the ride so I had him Export and chare his GPX file when I uploaded the same file that he used from the same device (his Garmin550) his elevation showed 2400ft while my upload of HIS file showed 1600ft. So, figure that one out. Almost like my Strava account is ripping me off ….. personally cause it hates me or something.
The positioning of the phone/device will play a big part in the accuracy of the data. The GPS technology itself in the iPhone 4S is a Broadcom BCM4750, whereas the Garmin seems to be using a SiRFstarIII. The implementation details (refresh rate of the data, the software algorythym used to smooth the data etc) will play a part in the accuracy also (trade off being battery life and other things).
For startup time, the iPhone is likely quicker as it can make use of AGPS to speed up the time it takes to get it’s first fix.
The Garmin is able to make use of a Barometric sensor to improve accuracy on big height changes (maybe relevant depending on where you cycle – or if you do downhill).
The newer iPhones (4S onwards) have GLONASS positioning as well as GPS, which will further improve accuracy (Garmin does not have this).
The Garmin is potentially using a speed sensor on the wheel to determine it’s speed and distance (Garmin prefers the data from the speed sensor over the GPS) and this could also be helping improve it’s accuracy – you can also fit a Speed/Cadence sensor and use it with the iPhone though.
All in all, for cycling, I think the positioning of the device and it’s view of the sky is likely to make the most difference – a phone in a pocket is at a disadvantage.
There shouldn’t really be cause of problems using an iPhone with Strava, my experience is that it performs fine – a speed sensor may improve things slightly.
Kinda what I was thinking. Not always the device in general or all iPhone 4s’s as a whole would perform as badly as this test shows. I started recording on an old Android HTC Evo (the original from 2010) and for about a year it performed flawlessly. About a month ago in Feb, 2016 is started acting like this iPhone 4s and the tracks were all jagged and going way off course, missing segments and the like. I think it may be due to installing additional apps and filling the 430MB of device storage to the max as that is the only thing that happened around this time ……… aside from falling off a car at 50mph. But the digitizer did not get any further cracks in it from this even than it had from falling off my lap onto the driveway. Phone still works. and a previously non working camera started working again after the 50mph flight off the top of the car.
Hilarious!