[Closed] How accurate is Strava?

It's not hugely accurate in many respects, but distance generally seems about right, so I'd say that's not far off.

Elevation can be very wide of the mark, not entirely sure if Strava works it out based on their own algorithms and the route points, or takes the data from the device, in which case anything using GPS altitude is likely to be well off.

Accurate if you are kom. Not very accurate if you are 2nd

Unless you measure it you wont know 🙂
Distance wise it seems to stack up if it follows the route correctly. Over that distance I generally get a lot of match ups with other recordings etc. I think it plots the points then checks distance. As it's measured by a Garmin 705 I would assume it's reasonably close.

I think over that distance even if it's 5% inaccurate you'll be nearer 35 than 25 miles 🙂

It does tend to 'snap to' segments - so if you ride the fire road beside a segment it'll move your track to the segment but generally for distance and elevation it's as accurate as your gps receiver so 'fairly' good really.

try uploading to Garmin connect for a comparison, maybe?

depends on the device that you are tracking with. iphone/android not brilliant. garmins are a bit better.

not entirely sure if Strava works it out based on their own algorithms and the route points, or takes the data from the device

I think it depends on whether your GPS has a barometric altimeter. If it does, then Strava will take the data directly from the device as it is deemed to be more accurate than elevation databases. If it doesn't, Strava will calculate your elevation based on their elevation databases.

If you get dodgy elevation data, there is an option to correct it by clicking on the question mark below the total elevation.

I find the elevation data on Strava, Garmin Connect & Mapmyride to be very similar but it's way out on Endomondo - that's using a GPS with a barometric altimeter.

Distance data always seems very close on all of 'em.

I think the distance works quite well with regard to the phone apps like Strava and Endomondo. We have seen some differences when doing the same loop with two phones both "recording" but they are quite small.

What doesn't work well is the cumulative climbing/descending stats. GPS location is calculation based upon satellite signals/locations and it has a little noise in it, this doesn't make much difference to distance but it makes a bigger difference to altitude. If you sit with a GPS device with an altitude read out you'll see it move around even if you stand still, the translates to exaggerated climbing/descending stats. Often this is visible as you ride a loop and the climbing/descending stats don't match.

More sophisticated GPS devices than phones have a signal strength / accuracy measure so that the user can estimate how accurate the position given is.

That person has done the lower loop twice though.
Knock that out it goes down to about 22miles.

That person has done the lower loop twice though.

Oh well 🙁

[i]He has?[/i]

zoom in on the map and you can see two tracks where gps inaccuracy hasn't followed same route. Looking at overall view it's tricky to see, though.

Just looked again and it looks as though he's done Monkey twice as well!

It get worse!

zoom in on the map and you can see two tracks where gps inaccuracy hasn't followed same route

Blimey, that ride's got more segments than a whole forest of satsuma trees!

I guess creating endless segments is the only way to get KOMs in civilisation (other than pedalling really fast). Round my way, you pick a hill, any hill, and the chances are you get the KOM...

IME, Strava is pretty inaccurate, but it's all a laugh innit? As per Van Halen, sometimes the inaccuracy works in your favour, sometimes it doesn't...

zoom in on the map and you can see two tracks where gps inaccuracy hasn't followed same route. Looking at overall view it's tricky to see, though.

Or just drag your mouse along the profile map to watch the 'rider' move along the map track, easy to spot then.

I think it depends on whether your GPS has a barometric altimeter. If it does, then Strava will take the data directly from the device as it is deemed to be more accurate than elevation databases. If it doesn't, Strava will calculate your elevation based on their elevation databases.

More or less, yes. Their algorithm is quite complicated.

https://strava.zendesk.com/entries/20965883-elevation-for-your-activity

I would think that for total distance, Strava is accurate within 1-2%. It sometimes has 'jumps' in timing which as stated can either improve or worsen your segment times significantly.

Strava does seem to have a tendency to cut corners. 😉

Just used Strava for the first time, on a 2 mile run from my house. I know it's 2 miles as I've measured it in Memory Map and with my bike computer and a garmin. Strava just told me I ran 3.6 miles in 15 minutes....

A lot of the rotue is under trees around the edge of the park, which might throw out the gps tracking. Also, when I look at the route in Strava, it is really "zig-zaggy", which is why the distance is nearly doubled. Strava doesnt seem to smooth this out, like Garmin Connect does.

swinleys new trails always show up 20% shorter than they should be on my strava I think it's the tree cover and lots of doubling back twisty turny trails

Sure you're not one if those corner cutters kimbers? 😉

I've just lost several KOMs to some suspiciously good riders around here. I mean, sure, they're pro riders on UCI world tour or continental tour teams but surely they can't be THAT much faster 😯

Perils of having local road rides on the same routes that a pro race happens on. Even the fairly handy local riders are looking a bit lack lustre.

There do seem to be timing inaccuracies depending on the device used too. On more than one occasion (in fact....lots of occasions) we have noticed discrepancies - such as when one of us trounces the other on a segment yet still comes out significantly slower.

The biggest errors seem to come with phones, they seem to be quite different to dedicated gps devices. I can only assume it is down to the sampling rate.

Distances though seem within a few % whatever you use.

fuzzhead seems to have got it right though.

milage is usally about right but climbing is way off ( or garmin is way off)
i ride with a chap whos got a garmin his climbing is allways a lot more on exactly the same route
did lakes last week i had 4800 feet on strava
mate had 5300 feet on garmin

I wouldn't say distance is accurate. A difference of 4km was recorded on a 30kmn ride between my mates Strava via I-phone and my Garmin device. Large difference on the total climb as well. Could be either or both out though.

not strava but iphone gps can be pretty accurate compared to a wheel speed sensor.

[url= http://farm6.staticflickr.com/5539/9040353251_40966d8a29.jp g" target="_blank">http://farm6.staticflickr.com/5539/9040353251_40966d8a29.jp g"/> [/img][/url]
[url= http://www.flickr.com/photos/jambo_uk/9040353251/ ]Test[/url] by [url= http://www.flickr.com/people/jambo_uk/ ]jambo-uk[/url], on Flickr

Strava struggles with loops; the Monkey has had several diversions over the last 12 months or so (due to trail repairs and tree-felling operations)...some of the diversions have been fairly significant but Strava can't differentiate.

[url= https://strava.zendesk.com/entries/20950148-Segment-Matching-Issues ]Some segments are frequently more challenging for our current segment matching algorithm. A segment that is a loop and has similar start and end points will not be able to distinguish direction, meaning whether it was performed clockwise or counterclockwise. Segments that include multiple loops through the start/end points will always only match one loop, no matter how many loops the segment actually is. This can cause for some very strange times as the time it took you to complete one loop is stretched over the actual segment distance which can include multiple loops.[/url]

Still love it though...it's great for measuring my own progress.

I deleted off my phone it be because of its inaccuracies, useless app if its inaccurate IMO

it be because of its inaccuracies

my climb rate is always higher on strava than garmin, according to strava ive climbed 10,496m in the past 28 days but garmin shows 10,122m.
im obviously going with strava 😀

I've just lost several KOMs to some suspiciously good riders around here. I mean, sure, they're pro riders on UCI world tour or continental tour teams but surely they can't be THAT much faster

Perils of having local road rides on the same routes that a pro race happens on. Even the fairly handy local riders are looking a bit lack lustre.

Without any data to look at you'll find that they probably ARE that much faster. Plus could have been full TT training or even motor pacing on the road.

swinleys new trails always show up 20% shorter than they should be on my strava I think it's the tree cover and lots of doubling back twisty turny trails

No, they are 1-2miles shorter than the signs say they are. I presumed it's because there's obvious scope to extend them in places like that random tabletop at the top of the first switchback climb on the blue (look left and it looks like they cleared a trail into it), W9Y, Red 4, etc, so the planned distance is longer than what was actualy built.

As for inacuracies, at worst my garmin seems to be 5-10m out (and for soem reason, consistent, i.e. 2 traces will be shadows of each other, just a few meters appart). So that's what, 1-2 seconds?

If you're bothered about <30second segments then that's a problem, but I only really keep an eye on the longer 5 minute+ ones as they're a more realistic guage of fitness/skill/balls anyway.

If your GPS track is rubbish then your Strava data will be rubbish. Most phones seem to have fairly poor GPS in them.

If your GPS track is rubbish then your Strava data will be rubbish. Most phones seem to have fairly poor GPS in them.

This. If I leave my phone in a pocket, shit GPS track.

Rear Jersey pocket on road bike or top pocket of my bag on the mtb and its as good as a garmin.