[Closed] how accurate is strava?

2 Levels, one is your GPS device, phones tend to be less accurate or lower smapling intervals than dedicated GPS devices.

Next is the segments and start/end points which goes on which side of the line your GPS last sampled.

So answer is it varies, just stick to it being fun for times and look at the data if you want to train to it

Not accurate enough to take it seriously, but good enough to wind your mates up😀

From experience there's always been a massive variation between different phones and devices, usually wildly inaccurate over short segments (i.e. those that take less than a few minutes) - you'll often get some very confusing results if you are trying to beat your mates time over that 25 second segment 🙂

Not very if using your phone. It had a 1k swing over the same 6.5k run around a local reservoir.

Seems better via my Garmin watch but tends to knock a bit of distance off the Garmin figure.

Pretty much.

It’s not even slightly accurate if you are trying to compare yourself to others.  If you are comparing your own performance with your other own performances it’s accurate enough to be a useful training tool.  It’s like pressure gauges or any other measuring device - stick to one device and you take a lot of variation out

Mostly the inaccuracies are with the recording devices - all dedicated hardware will always be more accurate than any phone but there are differences here still, basically it comes down to runtime (battery use) and methods of establishing location.  The ‘inaccurate’ phones are probably still fine for navigation if you’re ok with caning your battery that way.

I have a more or less standard training route. No two days data are the same. Doesn't matter if I use the phone app or my Bike computer. If I use them both at the same time, distance and height varies between the two. If I compare one day to another the same applies.

Just pick your preferred system and use that. At the end of the day it doesnt really matter unless you are a bit of an anorak.

Autopause functionality causes havoc.

Segments that start or end with hairpins can mark some runs starting having already negotiated the bend. Top of the leader board will be dominated by those with less accurate GPS.

Badly designed segments where the start point is too close to the point where people loiter at the start or linger at the finish can mis-trigger.

Short segments suffer disproportionately from the above

That depends.

If I'm fast it's blindingly accurate; if I'm slow that it's shockingly random

depends if anyone is using an ebike 😉

Top of the leader board will be dominated by those with less accurate GPS.

funny that, because top of the leaderboard round here is dominated by all the proper fast riders round here.

Really not enough to worry about.

Looking back over mine, I ride pretty much side by side with my Mate, we've both got iPhone 8s - generally one of us will ride 300-400 metres more than the other, climb 10-15 metres more than the other, sometimes it's me, sometimes it's him.

I remind myself of this when I'm trying to kill myself to shave a second off a sub-minute section.

However it markets itself, it's best suited for Road Riding or at least for sections measured in kilometres and not metres.

Also, the whole 'KOM thing' or just racing mates, it can be pretty toxic, I'm after smiles per hour, not miles per hour.

I tell myself I only care about yearly total KMs but I still get caught caring about Strava Segments, I'm going to switch to AllTrails in the new year, if nothing else I seem to have bought the Pro upgrade for £22 by mistake, I might as well use it.

One of the areas round here the leader board is dominated by people with well sub 10s times on steep twisty 200m trails. Knowing some of the people with KOM's and some of the really fast local riders times, there is no way the top end figures are accurate.

A mate I often ride with usually rides and extra km for every 25-30km ride we do according to strava. Both record with phones.

Not that accurate, and people who have bought GPS devices like to always assume that spending more money makes it better and more accurate, despite the fact that it is far easier to reduce the sampling rate on a stand alone GPS device...

Unless of course you're coming out on top on all the segments that day despite going in first and coming out last 😀

funny that, because top of the leaderboard round here is dominated by all the proper fast riders round here.

The original comment was specifically linked to the matter of start or finish sections doubling back in tight bends. Having shyte gps giving a 20 second advantage is something that fast riders can obtain as much as slow riders and the ticket is a shonky gps setup.

i'm not claiming strava is infallible.

but overall, fast people are fast on strava, slow people are slow.

IME it's the ones who think they are faster than they actually are who make the most noise about the inaccuracy of strava...

I dont know if accurate or not but used for the first time this weekend to the Pub (did not stop) the drive is normally 3.3 miles and data was exactly the same on Strava,  I was happy with that will try same route on weekend.

IME it’s the ones who think they are faster than they actually are who make the most noise about the inaccuracy of strava…

Of course - the segment where strava showed me doing 40mph+ up a steep climb, that was definitely accurate, it's only the one's which show me as 2 or 3 seconds behind my mate that i'm unhappy about...

From very extensive research, I've found it to be incredibly accurate when you are faster than anticipated. Conversely it is very inaccurate when you are slow, FACT!!!

If anyone disagrees they are spreading fake news.

Someone once said to me, "why do you want to race the internet? strava is for folk that dont actually race" :/ my opinion is its very inaccurate and makes folk think they are faster riders than they actually are, me included but generally faster guys are top of the table despite the inaccurate times, Ive strava'd races and compared times to official race times and the difference can be 20 odd seconds of a difference, sometimes more.End of the day strava is a bit of fun, nowt more.

I've tested with a couple of devices at the same time on pretty short segments (1 min) and you get differences of up to 5 seconds (phone vs element bolt) so as above, don't take it too serious and if you lose a KOM or someone is faster than you then its their GPS that is inaccurate, if its the other way round then yours in correct!

Most phones now have 1000ms sampling rate, which is good enough.

I've found if you record your own segments, and when doing so, go slowly to make sure there are enough data points then it seems to be pretty accurate. As above, anything under 1min should be ignored as they tend to be the least accurate.

I do a similar loop on my lunch break most days and distance and elevation are always within a few m, so it is consistent. Looping the same segment that I've recorded results in times within a few seconds, which is what you'd expect.

Strava isn't accurate enough to race with, but as long as you control as many of the variables as you can, does make for a useful tool.

but overall, fast people are fast on strava, slow people are slow.

Also this.There's always the odd outlier, but looking at segments as a broad spread this seems to be true.

Strava won't let you create segments shorter than about 300 metres now purely because of the inaccuracies involved.

For segments I tend to go off the basis that most of the time my PB will be 150% of the KOM's unless it's a downhill in which case I'll be double or more!

My main competitor is the dodgy looking fellow I see in the mirror each morning 🙂

My anecdote:

5 miles almost straight segment on a moor road (so good satellite sight  lines). Me & a mate with Garmin Edges. Start together, I finish 50-100 yards ahead. He beats me by 4 seconds. So not that accurate.  But given the variables of group riding, wind direction etc etc etc anyone who takes Strava leaderboards as anything but a bit of fun really needs to grow up.

Having your sampling rate as “Smart” (which is often the default) rather than “1 sec” has a massive difference.  Change it to 1 second sampling on a dedicated GPS device for best results.

From my experience, it’s pretty damn accurate 99% of the time. Especially for longer segments. It’s the short ones with bad GPS coverage that can be troublesome.

Most people have had a bad GPS ride and did 800mph up a hill and beat a 2 minute segment by 3 mins. Just ignore that ride that accounts for less than 1% of your Strava data. The rest of it is pretty accurate.

As already said, those that think they are faster than they actually are happen to make up most of the  “Strava is totally inaccurate bandwagon” and they’re probably chasing short downhill segments in the woods with Smart recoding turned on which kind of defeats the purpose, i.e. riding up hills.

The statement that dedicated gps is better than a phone is a myth

I'm sure some phones are rubbish. On dedicated devices cost doesn't seem to seem to help but size does

Just look at how much better is a iphone 5 is than a Fenix 5X

So inaccurate devices make for shorter segment times?  Anyone got any recommendations 😂

It's more than just hardware too. Conditions on the day (clouds) and fractional differences in your position and how that affects line of sight of satellites, e.g. trees and buildings. Angle of the device at the time even.

Many devices will report a lock with a few satellites fixed but better accuracy is achieved with a large or complete number of fixes. Also adding in Russian GLONASS may improve things. Assuming the Russians aren't nobbling it (or Americans likewise with GPS).

That's just position accuracy. Altitude/elevation varies on device, i.e. a device may use barometer reading (mostly in standalone devices, and not all of them, very few phones), GPS altitude (not very accurate at all), or GPS corrected to known elevation data in the case of Strava and some other sites. The latter is maybe more reliable but only down to a square area (10/50/100 metres etc). The ground could actually go up and down a fair bit within that square. Barometer is actual elevation read from pressure but to be accurate it needs to be calibrated based on known elevation at particular locations and adjusted for weather pressure changes (don't know any bike type GPS devices that can do this).

Distance accuracy is best if you use a wheel sensor but few do and you'll need a dedicated device with support for them also. These will track the actual distance covered by the wheels. Even then the sensor needs calibrating for wheel size, though many devices will attempt to auto calibrate. If you're into "air time" then it's less accurate when the wheels are in the air 😉

Consistency is key. Compare your own data and leave comparing with others as a curiosity and bit of fun.

IME it’s the ones who think they are faster than they actually are who make the most noise about the inaccuracy of strava…

Oh I don’t know, I’ve heard some holders of slightly improbable KOMs bleat about how amazing it is 😉

So inaccurate devices make for shorter segment times?  Anyone got any recommendations 😂

i've actually stopped using my Edge 1000 on some rides, its too accurate, theres always a certain local to beat in their backyard, and he uses the android app, so i use something less accurate, levels the playing field 🙂