strava time variations between devices

bianchimoon
bianchimoon Posts: 3,942
edited July 2014 in Road general
Interesting run the other night, 3 different devices, garmin 800, garmin 200 and strava iphone app
all riding together.
Segment 1 garmin 800 finishes 1.2mile sprint 1 bike length ahead of garmin 200
garmin 200 has a time 5 seconds faster
Segment 2 (4.36mins) garmin 800 and iphone start / finish at side by side iphone
iphone has 7 second better time than 800, garmin 200 1 second slower than 800
Must be some quiet big variations in the way the devices handle the timeline/updates to explain these differences.

Good job it's only a bike ride :D
All lies and jest..still a man hears what he wants to hear and disregards the rest....

Comments

  • buckmulligan
    buckmulligan Posts: 1,031
    Your experiences are most likely a results of the varying sampling/recording frequency between devices and just plain old luck as to where those points are recorded in relation to the start/end of the segments.

    This is well worth a read if you haven't done so previously:

    http://blog.veloviewer.com/41mph-the-ev ... s-article/
  • bianchimoon
    bianchimoon Posts: 3,942
    Your experiences are most likely a results of the varying sampling/recording frequency between devices and just plain old luck as to where those points are recorded in relation to the start/end of the segments.

    This is well worth a read if you haven't done so previously:

    http://blog.veloviewer.com/41mph-the-ev ... s-article/

    that explains it perfectly, thanks :)
    All lies and jest..still a man hears what he wants to hear and disregards the rest....
  • bianchimoon
    bianchimoon Posts: 3,942
    I was about to write a furious internet rant with the general message of "who cares?" but it seems you don't really care either, thank god! :wink:

    :D what you mean i don't get riches and fame for my fabled strava lantern rouge times.. damm i'll stick to the day job then :wink:
    All lies and jest..still a man hears what he wants to hear and disregards the rest....
  • bflk
    bflk Posts: 240
    I have a Garmin 200 and I am quite frustrated at times at how infrequent the 'smart' recording is - can be 15-20s between saved reads. Any time I am comparing myself with another Strava user the first thing I check is what device they were recording with, but I tend to ignore anything less than 3-4min long efforts. That eliminates an awful lot!
  • Buckie2k5
    Buckie2k5 Posts: 600
    why have you got smart recording on instead of every second?
  • rolf_f
    rolf_f Posts: 16,015
    My Brytons record every 4 seconds - mostly absolutely fine and it keeps the file sizes down but it maybe the reason why I tend to get a lot of identical fastest times for segments.

    Ultimately, mostly it is about your performance against yourself. Comparisons to others are fun but shouldn't be taken too seriously. Except for the Rapha climbing challenges where I tend to have to climb 25% further to match the Garmin users! :lol:
    Faster than a tent.......
  • BrandonA
    BrandonA Posts: 553
    Rolf F wrote:
    My Brytons record every 4 seconds - mostly absolutely fine and it keeps the file sizes down but it maybe the reason why I tend to get a lot of identical fastest times for segments.

    Ultimately, mostly it is about your performance against yourself. Comparisons to others are fun but shouldn't be taken too seriously. Except for the Rapha climbing challenges where I tend to have to climb 25% further to match the Garmin users! :lol:

    Depends what distance you are measuring performance against. Over a short sprint or hill climb your times on Strava will be wildly inaccurate due to the low recording rate of every 4 seconds.

    To put it into prospective. My Garmin records every 1 seconds. I took a TCX file from a ride and wrote a program that split this ride into 3 separate rides. One ride took the 1st, 4th,7th etc time recordings, another 2nd, 5th, 8th etc and the third 3rd, 6th and 9th etc. i.e. I made three inaccurate rides which tracked time every 3 seconds (which is what the Stava app does - or at least did do last year - not used it this year).

    I then uploaded the 3 rides to Strava. On one short downhill segment the times records on the segments were 26 seconds, 27 seconds and 30 seconds. The differences were due to when Strava found the start and end points - the lower the recording accuracy the more fluctuation you'll get in these figures. The original ride recorded a value of 27 seconds.

    If you ride along at 20 mph you are covering 8.9408 metres per second. To record an effort in Strava it needs to detect a trackpoint that is close to the start and end of the segment and use these to record duration beween the two points.

    If I start 2 Garmins half a second between one another they will be recording times differently. This means that Strava could think you start and end a segment 1 second out between the two segments. If you could -1 on start and +1 on end which would result in a 2 second difference. The more inaccurate your GPS is recording progress the larger the potential differences.

    The above is made even worst if your GPS is not recording accurately. Just because you are somewhere on planet earth, it does not mean that your GPS recorder agrees as they have a certain margin for error. This margin for error can result in Strava thinking you started/ended a segment at a different time than you actually did.
  • Dippydog2
    Dippydog2 Posts: 291
    Rolf F wrote:
    My Brytons record every 4 seconds - mostly absolutely fine and it keeps the file sizes down but it maybe the reason why I tend to get a lot of identical fastest times for segments.

    Ultimately, mostly it is about your performance against yourself. Comparisons to others are fun but shouldn't be taken too seriously. Except for the Rapha climbing challenges where I tend to have to climb 25% further to match the Garmin users! :lol:
    The best approach is to load your ride to Strava from your device. See how much elevation it gives you. Then press the "elevation correction". If the corrected elevation is more you stick with that. If it is less you delete the ride and reload from your device.

    simples. :wink:
  • rolf_f
    rolf_f Posts: 16,015
    BrandonA wrote:
    Rolf F wrote:
    My Brytons record every 4 seconds - mostly absolutely fine and it keeps the file sizes down but it maybe the reason why I tend to get a lot of identical fastest times for segments.

    Ultimately, mostly it is about your performance against yourself. Comparisons to others are fun but shouldn't be taken too seriously. Except for the Rapha climbing challenges where I tend to have to climb 25% further to match the Garmin users! :lol:

    Depends what distance you are measuring performance against. Over a short sprint or hill climb your times on Strava will be wildly inaccurate due to the low recording rate of every 4 seconds.

    One would hope that Strava would interpolate correctly - in theory, if your pace is constant for a couple of seconds or so before the segment starts and after it ends, then it should result in a fairly accurate time.

    I wish I could change the recording rate as an experiment but, tbh, I'm happy enough with it as it is.
    Faster than a tent.......
  • rolf_f
    rolf_f Posts: 16,015
    Dippydog2 wrote:
    Rolf F wrote:
    My Brytons record every 4 seconds - mostly absolutely fine and it keeps the file sizes down but it maybe the reason why I tend to get a lot of identical fastest times for segments.

    Ultimately, mostly it is about your performance against yourself. Comparisons to others are fun but shouldn't be taken too seriously. Except for the Rapha climbing challenges where I tend to have to climb 25% further to match the Garmin users! :lol:
    The best approach is to load your ride to Strava from your device. See how much elevation it gives you. Then press the "elevation correction". If the corrected elevation is more you stick with that. If it is less you delete the ride and reload from your device.

    simples. :wink:

    Devious!! :lol:
    Faster than a tent.......
  • guinea
    guinea Posts: 1,177
    My phones (Nexus 5 and iPhone 4s) both wildly overestimate the metres climbed during a ride.

    Great for completing the Rapha challenge, but it's all lies.

    My Garmin almost exactly matches the numbers expected from the big alpine climbs, but if I want a few extra hundred metres to boost my stats I use the iPhone on the work paid data plan :)
  • buckmulligan
    buckmulligan Posts: 1,031
    Dippydog2 wrote:
    The best approach is to load your ride to Strava from your device. See how much elevation it gives you. Then press the "elevation correction". If the corrected elevation is more you stick with that. If it is less you delete the ride and reload from your device.

    simples. :wink:

    I always thought that for most well-travelled roads, 'corrected' elevation data would be far more accurate than a single device's recording, especially if that device is using GPS-based elevation measurements, rather than say a barometric altimeter as is common in many Garmins.

    However, on the occasions that I've tried it on Strava the general elevation profile looks good and the elevation for the summits and valleys are in the right ball-park, but there's so much noise in the elevation data (little spikes up and down over the whole ride) that the cumulative elevation gain for a ride is much higher than it should be (20-35% IME). Not sure if that's always the case but I rarely bother trying it these days and just accept the data from my device, good-or-bad.
  • BrandonA
    BrandonA Posts: 553
    Rolf F wrote:
    BrandonA wrote:
    Rolf F wrote:
    My Brytons record every 4 seconds - mostly absolutely fine and it keeps the file sizes down but it maybe the reason why I tend to get a lot of identical fastest times for segments.

    Ultimately, mostly it is about your performance against yourself. Comparisons to others are fun but shouldn't be taken too seriously. Except for the Rapha climbing challenges where I tend to have to climb 25% further to match the Garmin users! :lol:

    Depends what distance you are measuring performance against. Over a short sprint or hill climb your times on Strava will be wildly inaccurate due to the low recording rate of every 4 seconds.

    One would hope that Strava would interpolate correctly - in theory, if your pace is constant for a couple of seconds or so before the segment starts and after it ends, then it should result in a fairly accurate time.

    I wish I could change the recording rate as an experiment but, tbh, I'm happy enough with it as it is.

    It doesn't as it's impossible for Strava to know what type of acceleration or de-acceleration took place immediately before/after the segment end/start - all it could do is predict an average which is just as inaccurate as not including it.

    For example, you might be sprinting into a segment and you may either tire or power after it. As I said, you move 8.9 metre a second at 20mph, with your device recording every 4 seconds if I were holding a constant speed of 20 mph that is 35.6 metres that is a huge distance and anything can happen during that time at a top of a hill.

    If you really care about Strava segment times then you need to get a GPS recorder that records position every second and records position accurately. Or you accept that the shorter the segment the more worthless/inaccurate the leaderboard positions are. On a full ride then recording every 3 or 4 seconds is fine as this will have little impact on the overall ride stats.