[Closed] Bike computer and Strava are different. Why?

Aliens?

It's probably aliens.

Different accuracy resolution, different algorithms to translate the data, number of satellites. Etcetc

But yeah, mostly aliens

.

A man with one GPS knows where he is. A man with two is never quite sure.

Another thing is how frequently it is recording trackpoints. eg if it only records a point every 10 seconds, it can cause it to cut corners.

Also, did you start them both at the same time? They could calculate paused time differently, which can make a difference for speed.

I'm running this app with work, for some charity deal the company pledges money for miles done or something. Anyway, I'm not very good at it, I always forget to start it off (why I've never bothered with Strava).

But when I did remember - I ride 9-10 miles home from work, look at the app "You done 7.3 miles" eh?

Then I go for a mtb jaunt around QECP, I think the red trail is 4.5 miles or something. Get to the end, "You ridded 9.5 miles". eh?

Aliens?

Ah. I see.

The truth is out there.

Somewhere!

If there's some "noise" at the beginning of the ride when the phone or computer is sorting the GPS signal out, that could be it.

Other points given could also apply, I guess the one with more frequent point taking or less accurate GPS would say you went further/climbed higher.

The climbing could be due to the cycle computer having a barometric altimeter that should be more accurate- if you don't have this Strava just projects your ride on a horizontal plane onto a rough mapping of elevation.

Because science.....

https://en.wikipedia.org/wiki/Error_analysis_for_the_Global_Positioning_System

If you download the XML of a GPX file you'll find it includes latitude and longitude, not distance, so exact placement of those points, as well as whether they do clever things like "snap" to roads will change distance and elevation.

Speed also varies because Strava does slightly weird things around stopped time - where trackpoints may equate to a speed of 10-5-0-0-5 a Garmin (for example) ignores the 2, and comes up with an average of 6.7, whilst Strava ignores the first zero and 'draws a line' between the last point where a speed was recorded and the the final zero, and would make the first zero a 2.5mph ie it doesn't accept a short period of 0mph, it thinks you slowed down very progressively, stopped instantaneously and then moved off. At least it did a few years ago when I looked into this becuase it annoyed me why Strava was always slower! This is obviously exaggerated the more you stop.

What I don't know is whether there's an "outside bound" to that where it accepts that you really were stopped.

My watch always seems to rob me of .3 of a mile or so at park run , everyone else seems to run 3.2 miles but I’m always 2.9 or so . Tomtom and my new garmin does the same .

My experience is that the website treats the data differently, although distance, average and maximum speeds are pretty close across sites using the exact same GPS file synced across all of them

Altitude varies a lot though, showing they all use varying levels of data interpretation and smoothing.

My last 3 identical rides, recorded on a Polar GPS watch, synced to Polar website, and then synced to Strava and Endomondo...

Total Distance on Strava is about 200m less per 100km than the others (I can live with that level of accuracy)

Max Speed identical across all 3. Average Speed Strava gives me 1km/h more.

Total ascent varies wildly. Endomondo short changes by about 30%. Strava gives you about 40% extra free.

So use many services, and choose whichever number is best 🙂 😉

If it were recorded on different devices, the total differences is only going to be larger.

A man with one GPS knows where he is. A man with two is never quite sure.

Damn, that's a fine quote. I may steal it, if you don't mind.

Strava Tax, innit?

They take a few metres off each freeloading Strava user and add it on to those that pay the premium Strava subscription. Only fair, really.

Rachel

GPS gets more battery intensive as resolution / accuracy goes up so most bike/wearable models are not terribly accurate at rapidly moving objects but would be fine at walking pace.  Packaging dictates small battery which in turn dictates less frequent satellite pings which makes the ‘accuracy’ of some quite laughable when split seconds are quoted. Mobile phones often mix in a bit of cell mast triangulation where available as it’s cheaper in battery terms and also drops accuracy noticeably.

Rotating wheel counters can be very accurate but need careful calibration for actual wheel size (‘700c’ etc is not the right answer here... 😉)

Probably neither is terribly accurate.  Maybe plot both and draw a line of best fit? (Joke!)

Thanks guys. It's all a bit of a mish mash isn't it?

Strava and computer give different segment results so I could keep both and improve my performance remarkably!

In practice I think I'll save battery power on the phone and just use the computer. It may not be totally accurate - but what is - but it will be consistent.

Just to add a further complication.....if I'm doing a totally new route then I will plot this into my Garmin eEtrex 30 GPS and use that for navigation. That then gives me a track which will give yet another different set of results. I'll  just pick the one that suits me best

Which iGPSport have you got OP?

I've just got the i10. Good bit of kit so far and pairs to my mish-mash of Ant sensors well. 👍

it will be consistent

That's the main thing to go for. None of this stuff is accurate and there are so many variables. If it's vaguely consistent then that's good enough.

Gives me a comparison to previous rides for performance and distance, even though the distance figures are never perfect (I wonder about going back to a wheel speed sensor like I used to use, to track actual ground rolled, not satellite distance, but wasn't convinced by that either). Elevation is the worst. Without a barometer it's either satellite calculation which is very poor, or based on ground data at locations which is limited in resolution. With a barometer, weather pressure conditions during a ride mess with it. If I start and end in the same place I can be sure it will show a sizeable difference. Has been off by 100ft at times.

Go to another device, even the same make and model, and consistency is lost.

@poopscoop

Sorry, haven't worked out how the quote thing works!!!

I've got the iGPSport50E. Was cheap on Amazon "Only 1 in stock" or so they said. I had no real idea what I was buying, but didn't want to spend some of the eye watering amounts that Garmin demand. I was just after a wireless computer. All the Strava add ons were more than I was looking for. I'm using them because they exist, but for how long???

Phone uses data and GPS, bike computer uses gps and some have a barometric altimeter to measure altitude. I often have to login to strava and hit 'correct altitude' and distance comparing my mates who use phones to log their rides.

I always got my best times with a iphone...They seem the most "generous" (or innaccurate if you must)

I've uploaded half a dozen rides now, and the phone does seem to give me a better performance. Looking at the map I'm not sure that the altitude gain is right. One of my regular rides "recorded" a 50% jump in climbing then next day it went back to normal. That was on the phone. The computer was pretty consistent.

My mate and I were out road riding at the weekend. Rode together, exactly the same route, both using Wahoo Elemnts, but got different ride lengths and ascent figures at the end. Go figure.

So if you get a KOM by less than 5% its probably not really a KOM, its an error 🙂

No. It's only an error when someone beats your own KOM by less than 5%.

Will never happen me because I'll never get a KOM of my own.

Your phone was drafting the computer & your bum is 0.18 of a mile wide?

So if you get a KOM by less than 5% its probably not really a KOM, its an error

Any segment of less than 1 minute is a bit meaningless as GPS vagaries will have a significant impact. Obviously that diminishes the longer the segment gets. I'm not sure you can put 5% on a 30 minute segment down to error 😉

It's the ascent figures that bug me the most, trying to train for a ride with 5300m ascent, thought I'd put in a couple of rides at well over 2000m ascent until I actually checked the contours and found they came up well short.

Wondering now if the 5300m figure is correct for the original ride I was training for, if they've just taken that off a route planner it's probably wrong, and if they've taken it off their GPS at the end of the ride, it's probably wrong. Will check elevation profile and do a rough count I guess.

Also makes me wonder if the people who are apparently cracking out 1000s of metres elevation every ride are basing that on wildly innaccurate GPs/route planner elevations. That's what I tell myself anyway ; )

Edit: Hmm, as I thought, the 5300m elevation route is more like 3000m if you count peaks and troughs. Seems to be the usual margin for error on these rides, the route where I thought I had done 2900m was more like 1800m.

Altitude is something consumer GPS units have a problem with: head out for a group ride and the distances will be very similar, to within a couple of hundred metres, but there can be a 100% difference in the recorded altitude gain. Even riding the same route again with the same unit and there can be a big difference. Looking back at my recent commutes there's at least a 10% difference in recorded altitude gain but the distance is consistent to within 10 metres or so for a 21km ride.

Sometimes the GPS unit throws a wobbly - I do the monthly Strava climbing challenges and one month a guy in Singapore was the leader, he was managing 100,000m of climbing on a 10km commute! Looking at one of his rides there'd be a "blip" where he'd go from sea level to 10,000 metres in the space of a few metres distance then back again.

I wouldn't necessarily say it's the consumer GPS unit that has a problem with altitude. It's more an issue with whatever service is interpreting the recorded altitude data.

If I do a big loop out, starting and ending in the same place, then the absolute altitude at the start+end will more often than not report the same value +/-10m.

As per my previous post, whatever interprets the data (the Garmin, Strava, Endomondo, etc.) interprets it slightly differently. Some smooth out the data more than others. Hence why I got 326m, 420m and 660m altitude gain not only on the same ride, but all recorded on a single device, with the same GPX/TCX file synced to all 3 services. At least 2 must be wrong.

I will take the GPX some time and add up all the accumulated differences between every consecutive pair of data points.  Last time I did this, Endomondo was quite harsh at filtering out the slight variations, Everytrail was exactly the same as the accumulated differences, and Strava either made up altitude that was not in the file or has some kind of prediction or extrapolation that overshoots (or is trying to correct it according to a contour map or something).

I will also do a 100km ride in Holland some time to compare. I was also going to leave the GPS device sat in a fixed location for 2 hours, just to see how far it travels in distance and altitude. Neither will be 0 for sure.

I've had turbo sessions record altitude gain! I didn't think I bounced around on the bike that much.

If you are on a bike or on foot there's a maximum altitude difference possible between each GPS signal - you can't climb/descend 100 metres in 1.5 seconds for example - so I'd be filtering that data out.

Strava have an "elevation FAQ" -  https://support.strava.com/hc/en-us/articles/115001294564-Elevation-on-Strava-FAQs

Strava sees the world a bit like minecraft with 30 or 90 metres wide 'blocks' I think. Your gps device also has various sources of error. I suggest turning both off and enjoying the ride!

Hmmm... the force was strong with me on this ride...

Premier Icon
darrenspink

Subscriber
Phone uses data and GPS, bike computer uses gps and some have a barometric altimeter to measure altitude. I often have to login to strava and hit ‘correct altitude’ and distance comparing my mates who use phones to log their rides.

This.

Strave uses map data and won't log small changes in altitude where the map altitude is of a fixed value. You could cycle along a line that the map data says is 200ft but if it's not flat then your GPS unit using an baromaetric altimeter will pick up these changes and this is where you will start to see discrepancies in data.