I am an enthusiastic user of Strava. I suspect that friends I cycle with call me obsessive. They may even shout "strava whore" as I accelerate away in pursuit of a PB or to regain a KOM, but their words are, inevitably, lost in the wind. Strava has made me competitive. Fiercely and unashamedly competitive.
But all is not quite what it seems and recent events have conspired to make me question the veracity of all my Strava times and, more worryingly, the legitimacy of my single remaining KOM (shared I should add).
I use the Strava iphone app, but as battery life does not accommodate the demands of GPS on longer rides I borrowed a Garmin 500 from a cyclist friend for a recent sportive. I have had an opportunity to use the Garmin again and, purely out of my own curiosity, completed a ride using both the iphone and the Edge - using the GPS only on both devices and disabling the wheel sensor on the Garmin.
Whilst much of the data recorded by the two devices were exactly the same (or at least very close) some of the figures differed, most notably in the total climb and the segment times. The latter, of course, has a huge bearing on Strava leaderboard placings. These are a selection of the readings.
Device iphone Garmin Edge 500
Total distance 56.1 57.1
Total climb 1345 ft 948 ft
Time 02:49:29 02:55:42
Max Speed 40.7 mh 40.3mi/h
Av. Speed 19.8 mph 19.5mi/h
Segment I 00.39 00.35
Segment II 5.42 5.41
Segment II 2.00 1.59
Segment IV 5.56 5.57
Segment V 13.30 13.30
This selection of figures from the ride illustrates a couple of things: firstly, that the gradients are way out! Secondly, that the segment times fluctuate by about 1 to 4 seconds, irrespective of distance or whether it is a climb, a descent or pan flat. Sometimes the Garmin records a slightly faster time, but more often than not the iphone records the faster time. I would say, roughly that 45% of the recordings are faster on the iphone, 30% on the Garmin and the remaining 25% are exactly the same.
Armed with these stats, I approached Strava to see if they could enlighten me on the anomalies of the data. Their response was swift and illuminating. Firstly, the 400 ft disparity in the total climb over the course of a 56 mile ride. Their explanation was straightforward and even a technophobe like me could understand:
'The Garmin 500 has an internal barometric altimeter which measures elevation data. The iPhone does not, so we reference a known elevation database and "look up" elevation based on GPS coordinates. The two methods can sometimes yield different results, either because of the database used, or due to different smoothing factors.
It is pretty difficult to get the two completely separate data collection methods to yield the same results.'
Very difficult, if the discrepancies in the gradient recorded above are anything to go by! So, the fact that the Garmin has Barometric Altimeters means that it is more accurate, correct? Again, Strava’s response:
In general, the elevation collected by Barometric Altimeters is more consistent, and more "accurate". But mostly accurate RELATIVE to previous elevations. The Barometric Altimeter does not do ABSOLUTE elevation readings that well. But as far as elevation gain is concerned, only the relative elevation readings are important, so for that it can be more accurate than looking up values in an elevation database.
Personally, I take that as a 'Yes' - if you want accurate elevation gains and total elevation climbed, go with the Garmin.
The differences in the segment times require a little more explanation and understanding of the technology being used by the respective devices.
'Segments are matched based on your GPS coordinates, and which coordinates "best" match the coordinates of the segment. Two different GPS datasets will produce different segment times, as the location of the "best" matching GPS coordinates won't always line up perfectly with the segment. So, because the iPhone has a 3 second recording interval, and the Garmin has a more frequent recording interval, the segment times with both the devices could be different, especially on descents when the speeds are higher.'
'In short, the more points you record, the more accurate your data set is. So users who have 1-second recording find more accurate segment times than users with Smart Recording, or 3-second recording intervals. '
So this explains the differences in the segment times recorded and why sometimes the iphone gives a faster time, sometimes the Garmin. And if you can get your head around 'recording intervals' and 'segment endpoints', it becomes clear that the Garmin is going to be more accurate than iphones and all of Strava's mobile apps as they use a 3-second recording interval.
'When we match a segment, we look for the endpoints in your data that are "closest" to the segment endpoints. This could be either before or after the actual segment endpoints. With 3-second recording, there is a much less probability that your data will line up exactly with the segment data, as there are fewer points. There is a greater probability that your chosen endpoints will be inside the segment endpoints, giving you a faster time on the segment.'
On whether I should buy a Garmin to achieve a greater degree of accuracy, Strava cannot be drawn - but their explanation of the differences in the data suggest this is very much the case. Strava, and you have to give them a little bit of slack here, do not and will not recommend one system over the other!
So where does this leave me and other Strava iphone app users? Yes, we hold an unfair advantage over our Garmin brothers and sisters and, more importantly, it does bring into question all our times on the segment leaderboards.
So, in the highly unlikely event that I regain the King of the Mountain that was so cruelly taken from me recently, I will now think twice before I hold the severed head of my usurper aloft, metaphorically speaking of course. Instead I will check if my vanquished foe was using an iphone app or a Garmin device. I would hate to be accused of being a Pretender to the Throne.
Strava's answers to my questions have left me glancing at my meagre collection of kudos (8) with a touch of shame and embarrassment.
But then again, all's fair in love and war! Isn't it?
Incidentally, Strava provided this useful link that sheds more information on how Garmin devices can record data.
Add a Comment