Support road.cc

Like this site? Help us to make it better.

Discrepancies between Cyclemeter and Strava Distances

Apologies if this has already been covered elsewhere, but I'm really struggling with this.

I have been using Cyclemeter (Elite version) on my iPhone to record my rides for a few months now, then uploading them to Strava once I've got back home. I then noticed a couple of weeks ago that the distance covered according to cyclemeter was significantly higher than on Strava once I'd uploaded it. For instance, yesterday, I went for a short 16-mile spin. At the end of the ride, cyclemeter reckoned I'd covered a smidge over 20 miles, and when I uploaded it to Strava it said I'd covered 16.3 miles.

I presume that cyclemeter is using the input from my speed sensor rather than the GPS of my iPhone. I've got a Panobike Cadence and Speed sensor on the bike, and I've put in my wheel circumference into cyclemeter (2099mm) which I'm pretty sure is correct - 700c rims with 622-23c tires. I don't think that there's any other settings that need to be configured.

When I look at the detail of my speed during the ride on cyclemeter, I top out at 40mph (which I think is frankly unrealistic) and on Strava it says it's about 33mph (which seems a lot closer to the truth  1 although I suppose I could be doing myself a disservice  4 ).

So I'm confused as to what's going wrong here. Any help would be much appreciated.

Thanks

If you're new please join in and if you have questions pop them below and the forum regulars will answer as best we can.

Add new comment

27 comments

Avatar
the_jm | 9 years ago
0 likes

It should be pretty easy to check. Get on the turbo, top gear, pedal up to 100 rpm and see what speed. Assuming 50 on the front and 11 on the rear you should see around 35.8mph. For any other ratio 7.87 * front/rear will give you a good enough approximation of the speed. If that is correct, you need to make sure cadence is being counted correctly. Count rotations for a measured time to check that.
It could also be poorly placed magnets, the magnet on the crank may be too close to the speed sensor, or the magnet on the wheel might be too close to the cadence sensor. (Edit: Looking at the differences you are getting I would say this could well be the problem. Your cadence magnet could also be registering on your speed sensor, try adjusting the position of the magnet realtive to the sensor)

Avatar
penguin116 replied to the_jm | 9 years ago
0 likes

Thanks for your advice the_jm. I'll have a crack at this over the weekend and see what happens.

Avatar
penguin116 replied to penguin116 | 9 years ago
0 likes

Success!  1 At least I think so. Looks like the cadence magnet was the source of interference on the speed sensor.

A quick 15 mile blast (according to Cyclemeter) with top speed of 30.2mph uploaded into Strava and it's come back with 14.75 miles and top speed of 30.0mph.

Hoping to go for a longer ride tomorrow, so we'll see if this extrapolates to wider inaccuracies, but I'm pretty happy with the change.

Avatar
the_jm replied to penguin116 | 9 years ago
0 likes

I typically find about 2% difference in distances, which is pretty consistent. A 100km ride registered as 98 on Strava and a 103 miles registered as 101. I have two of these sensors on different bikes and the results are near enough identical. If you're getting speeds and distances that are between 25%-35% out the odds are it is the cadence magnet that is the cause.

Avatar
giobox | 9 years ago
0 likes

GPS, while often very good at giving you an x,y coordinate, is really bad at calculating altitude. If you upload a ride from the Strava app, Strava uses map data for the x,y point of your ride to get the altitude, completely ignoring the GPS. This generally results in way more accurate elevation for a given ride.

If you don't use the Strava app and use your own dedicated GPS, this is what the "correct elevavation" button in Strava does - replaces your GPS unit's altitude numbers with values from their map data. If you have a good quality GPS it will have an altimeter built in to calculate altitude instead of using the GPS value. This is what the Garmin Edge 800/810/1000 do for example.

Avatar
fenix | 9 years ago
0 likes

2099 sounds about right - but what PSI and weight are you ? If its low - then surely that would reduce the figure and increase your milage ?

I'd put in a silly figure - half or double it to see if that is where the reading is being calculated.

Strava is accurate for me - it always compares well to my Garmin anyway.

So I'd go with that in future.

Avatar
Tintow | 9 years ago
0 likes

I use Cyclemeter and Strava with a Wahoo Blue SC and the readings are always pretty similar - never identical due to the issues described above but very close.

Avatar
penguin116 replied to Tintow | 9 years ago
0 likes

Abvio suggested that the Wahoo was the better sensor to use, but being tight-fisted (and resident in North Yorkshire  4 ) I went for the cheaper option. Should have shelled out the extra 15 quid (or whatever the difference was) and opted for the Wahoo.  2

I'll play with the settings for the wheel diameter and see if I can get a more accurate approximation for my speed and distance.

Avatar
the_jm replied to penguin116 | 9 years ago
0 likes
penguin116 wrote:

Abvio suggested that the Wahoo was the better sensor to use, but being tight-fisted (and resident in North Yorkshire  4 ) I went for the cheaper option. Should have shelled out the extra 15 quid (or whatever the difference was) and opted for the Wahoo.  2

I'll play with the settings for the wheel diameter and see if I can get a more accurate approximation for my speed and distance.

I too am from North Yorkshire and tight-fisted, hence I have the Panobike sensor too. The speed sensor seems pretty accurate with mine, riding in a bunch with others with various sensors and we all seem to be reported within +- 0.2kph of each other. It could be a QC issue though. I also do not get such big differences between Strava and Cyclemeter, Strava is pretty much consistently about 2% less than Cyclemeter records. This is annoying when you've just completed a metric or imperial century and Strava says not!

Avatar
penguin116 replied to the_jm | 9 years ago
0 likes

Aha. It didn't occur to me that it could be the sensor that's actually faulty. Maybe it's time to use Wiggle's 365 days returns policy.

Avatar
Man of Lard | 9 years ago
0 likes

Even if you were at sea, it would depend on the state of the tide.

Avatar
TimC340 | 9 years ago
0 likes

Speed/cadence sensors aren't particularly accurate devices; they're simple magnetic switches which can have errors from all sorts of reasons - wrong gap. too wet, battery expired (causing transmission errors to the head unit), etc. Strava will ignore the part of the .tcx file including the speed sensor info, and in this case will use the far more accurate GPS info. Not that GPS is infallible - and GPS altitude information from cheap receivers such as those on bikes is very poor.

The Bryton/Strava rollers session problem is a pain - I agree that if you've done an hour at FTP you should be able to analyse all the data, including speed and thus distance. I use a Garmin head, upload to Connect which then distributes to Strava. Alternatively, I use TrainerRoads (via an ANT+ USB dongle into the 'puter), download the session and then upload it to Strava. Either way seems to work ok.

Avatar
stealth | 9 years ago
0 likes

I uploaded my Bryton onto Strava yesterday & my 36.7mile turbo session from yesterday morning turned into a 0.3km ride at 0.2km/h average.

Avatar
le Bidon replied to stealth | 9 years ago
0 likes

Seems that even Strava knows that turbo sessions shouldn't contribute to your yearly total!  3

Avatar
andyp replied to le Bidon | 9 years ago
0 likes
le Bidon wrote:

turbo sessions shouldn't contribute to your yearly total!  3

This is quite literally the strangest cycling 'rule' of all time. Freewheeling down a big hill counts towards your yearly total, but an hour at FTP doesn't?

Avatar
fenix replied to andyp | 9 years ago
0 likes
andyp wrote:
le Bidon wrote:

turbo sessions shouldn't contribute to your yearly total!  3

This is quite literally the strangest cycling 'rule' of all time. Freewheeling down a big hill counts towards your yearly total, but an hour at FTP doesn't?

Not that strange. I just record it as a session - but its not miles.

I could set the resistance down low and generate lots of junk miles. Its a great training session - but its different to miles. usually better in fact - so keep the two seperate.

Avatar
le Bidon replied to fenix | 9 years ago
0 likes
fenix wrote:
andyp wrote:
le Bidon wrote:

turbo sessions shouldn't contribute to your yearly total!  3

This is quite literally the strangest cycling 'rule' of all time. Freewheeling down a big hill counts towards your yearly total, but an hour at FTP doesn't?

Not that strange. I just record it as a session - but its not miles.

I could set the resistance down low and generate lots of junk miles. Its a great training session - but its different to miles. usually better in fact - so keep the two seperate.

Exactly! Whether the "miles" are better or worse they are different and ought to be kept separately.

Avatar
andyp replied to fenix | 9 years ago
0 likes
fenix wrote:

Not that strange. I just record it as a session - but its not miles.

I could set the resistance down low and generate lots of junk miles.

Low resistance? Like...bimbling around a flat route in zone 1. Or a big downhill. Also 'junk miles', presumably, and stripped from your mileage?

Avatar
gdmor10 | 9 years ago
0 likes

from what I understand GPS doesn't do altitude very well, I remember being out sailing once and the GPS on the yacht said we were 3m below sea level - am pretty sure that was not correct.

Avatar
andyp replied to gdmor10 | 9 years ago
0 likes
gdmor10 wrote:

from what I understand GPS doesn't do altitude very well, I remember being out sailing once and the GPS on the yacht said we were 3m below sea level - am pretty sure that was not correct.

Depends if you were on the sea or not. Otherwise it's quite possible.

Avatar
Man of Lard | 9 years ago
0 likes

You'd like to think so... But logic goes out of the window when anything to do with computers is involved  1

Avatar
penguin116 replied to Man of Lard | 9 years ago
0 likes

Yes. I agree.  1

I've noticed on a couple of other forums that others are also seeing similar discrepancies not just with Cyclemeter, but a variety of different apps that do the same thing. The general suggestion is that it's due to inaccuracies in the GPS module of the smartphone (doesn't seem to matter if it's iOS or Android). I'm not convinced though - I'm still pretty sure it's something to do with the speed sensor taking precedence.

I have emailed Abvio support to see if they can shed any light on the issue, but as yet I've not heard anything in repsonse.

Avatar
penguin116 replied to penguin116 | 9 years ago
0 likes

I got a response back from Abvio. They said...

Quote:

Thanks for writing to us about your Cyclemeter vs Strava stats. I'm glad to help and I have a 2-part answer for you.

-- The discrepancies you are seeing are not uncommon. When you send your Cyclemeter data to Strava, they read in our data but then they use their own algorithms to calculate the values for things like distance, speed, stopped time and moving time. As a matter of fact, if you are using a speed sensor with Cyclemeter to determine your distance and speed, those values are not used at all by Strava. Instead, they use only your GPS data to come up with your distance and speed. So, again, seeing differences is not only usual, it's to be expected.

-- We don't have a Panobike Speed and Cadence Sensor in-house, so we haven't tested it ourselves. However, we have
had multiple customers write in with problems because the sensor was generating exceptionally high, invalid speeds. That's why we haven't added it to our official list of supported devices. Here's a page on our website with a list of devices that we do support:

https://abvio.com/question/devices/

Sorry I have answers - but no solutions. When/if you are in the market for a new speed/cadence sensor, we recommend Wahoo. We work very closely with them and our apps are designed to work with their products. However, there are other sensors that also will work well and you can check them out on our web page.

So this all suggests the inaccuracies are in using the Panobike speed and cadence sensor - anyone else had any problems with this?

Avatar
Man of Lard | 9 years ago
0 likes

Have you allowed the real GPS position on your iDevice? If you don't then the route it plots will be based on cell tower info and in a rural area that can be miles wrong (positionally)

https://strava.zendesk.com/entries/22533685-Strava-iPhone-app-and-new-iO... (this for the settings you need to look at)

Avatar
penguin116 replied to Man of Lard | 9 years ago
0 likes
Man of Lard wrote:

Have you allowed the real GPS position on your iDevice? If you don't then the route it plots will be based on cell tower info and in a rural area that can be miles wrong (positionally)

Ok. I've got that. It's all correct for Strava. Cyclemeter doesn't appear in the list, though. The general setting for Background App Refresh is enabled, so presumably if an app doesn't have a specific setting, the general setting would take effect.

Avatar
OnTheRopes | 9 years ago
0 likes

I cannot speak for Cyclemeter but when I have used the Strava App on my phone and also uploaded to Strava from the Garmin there has always been a big discrepency in altitude. I mean a BIG discrepancy like up to twice the amount on occasion. No idea why.

Avatar
Man of Lard replied to OnTheRopes | 9 years ago
0 likes
OnTheRopes wrote:

I cannot speak for Cyclemeter but when I have used the Strava App on my phone and also uploaded to Strava from the Garmin there has always been a big discrepency in altitude. I mean a BIG discrepancy like up to twice the amount on occasion. No idea why.

Because they use different criteria to decide what is a real altitude change (and what is GPS calculation error)

Latest Comments