jefmcg
Guru
Turn it off then. You will get the same speed if you are never below 2mph.
You may have the same problem - I've even given you a technical reason for it. But turn it off and check.
Is Garmin/GPS distance measurement inaccurate?
- Thread starter I like Skol
- Start date
Page may contain affiliate links. Please see terms for details.
You may have the same problem - I've even given you a technical reason for it. But turn it off and check.
OP
OP
I like Skol
A Minging Manc...
- Location
- Sunny Ashton-under-Lyne
This might be required. One of my calibration trips is from my house to Sid's cafe in Holmfirth. I can't remember the exact distance because it has been a while but it is close to 16 miles each way, so a there and back trip of 32 miles. Perhaps it is time I did this when running on GPS to see what it registers?
OP
OP
I like Skol
A Minging Manc...
- Location
- Sunny Ashton-under-Lyne
Do you even understand how GPS works? The reason we set a lowest speed is to remove the deviations in reported position that is inherent in all GPS position calculations as a result of calculation errors and signal fluctuation. If I want journey average speed I will turn off auto pause. Because I want moving average speed (I don't want to count the time I spend behind a bush having a pee or in a cafe stuffing my face with cake!) I select auto pause. The fact I have auto pause switched on means I am not counting the irregular position discrepancies that are inherent in the system as distance travelled when I am actually motionless. The problem I am experiencing is an under reporting of the distance travelled, and seeing as I probably spend less than 30m of a typical 10 mile ride moving at less than 2mph that 2.5% is being lost elsewhere.
Shut Up Legs
Down Under Member
Same here: I've found the Garmin GPS accuracy to be very good, and as @ColinJ said, the inaccuracy due to elevation not being included is very small. Now if only its barometric altimeter wasn't so subject to the vagaries of weather.
DaveReading
Don't suffer fools gladly (must try harder!)
- Location
- Reading, obvs
No. The quoted 70m height difference would need a barometer change of about 8 millibars. While you might get that, at a pinch, from day to day, it's highly unlikely to happen while you're sitting in the cafe eating your cake.
The Haversine Formula is required when distance calculations need to take into account the curvature of the Earth. For the average bike ride, it's completely irrelevant.
I have found GPS to be very accurate - on tour I always used to calibrate my computer on the first day against the road marker posts on long straight roads - normally about 10k (this is Europe not UK - UK marker posts are not to be trusted) - following that typically the GPS and the computer will agree to less than 1% - these days I tend to calibrate my computor against the GPS with the same result
Vertical - I have always found my GPS/barometer to be pretty good considering the inherent inaccuracies with GPS and vertical calculation - topping out cols typically less than 5m in it
Not withstanding this very rarely the GPS signal will drop out and according to the GPS I will have gone straight on - last year I cycled across the Rhine without getting wet- rare though that I can remember the two or three time it has happened
Vertical - I have always found my GPS/barometer to be pretty good considering the inherent inaccuracies with GPS and vertical calculation - topping out cols typically less than 5m in it
Not withstanding this very rarely the GPS signal will drop out and according to the GPS I will have gone straight on - last year I cycled across the Rhine without getting wet- rare though that I can remember the two or three time it has happened
Dogtrousers
Kilometre nibbler
There's a big difference between accuracy and precision. Here are a couple of links: NC State University ; Wikipedia.
From the OP: "[cycle computer] is repeatable and comparable results over a regular route (my commute which never changes) to within 0.5%" So we know that the cycle computer is precise. It gives reliably repeatable results.
From the OP: "[GPS] distances are consistently recorded around 2.5-3.0% lower". The key word is consistently. We know that the GPS is also precise. It gives reliably repeatable results to within about 0.5%.
So, we know that both methods are precise to within ~0.5%. But we don't know if either is accurate. For that we'd need a gold standard reference measurement. And we'd have to ask all kinds of philosophical questions about what we mean by distance and crucially why we need such accuracy, and exactly how much accuracy do we need for that purpose. And that would get very boring.
It's certainly not justifiable to say that the GPS has a "piss-poor" is inaccuracy of 2.5%. It could be that the holy grail reference measurement would lie slap between the two, in which case both would have a systematic inaccuracy of ~1.25%. It could be that the GPS is bang on and the computer has a systematic inaccuracy of 2.5%. How would you go about getting your reference measurement? Surveyor's chains? They will have a systematic inaccuracy too.
As we don't have a reference measurement, the best thing to do is pick one measurement system and stick with it.
This problem gets worse the more complex the measurement system is, and the more more likely it is to contain systematic bias. This is why we end up having this discussion more often about elevation and derived measures like "moving speed" and calories burned than about distance. But the principle is the same.
Pick one system and stick with it. What's the worst that could happen?
From the OP: "[cycle computer] is repeatable and comparable results over a regular route (my commute which never changes) to within 0.5%" So we know that the cycle computer is precise. It gives reliably repeatable results.
From the OP: "[GPS] distances are consistently recorded around 2.5-3.0% lower". The key word is consistently. We know that the GPS is also precise. It gives reliably repeatable results to within about 0.5%.
So, we know that both methods are precise to within ~0.5%. But we don't know if either is accurate. For that we'd need a gold standard reference measurement. And we'd have to ask all kinds of philosophical questions about what we mean by distance and crucially why we need such accuracy, and exactly how much accuracy do we need for that purpose. And that would get very boring.
It's certainly not justifiable to say that the GPS has a "piss-poor" is inaccuracy of 2.5%. It could be that the holy grail reference measurement would lie slap between the two, in which case both would have a systematic inaccuracy of ~1.25%. It could be that the GPS is bang on and the computer has a systematic inaccuracy of 2.5%. How would you go about getting your reference measurement? Surveyor's chains? They will have a systematic inaccuracy too.
As we don't have a reference measurement, the best thing to do is pick one measurement system and stick with it.
This problem gets worse the more complex the measurement system is, and the more more likely it is to contain systematic bias. This is why we end up having this discussion more often about elevation and derived measures like "moving speed" and calories burned than about distance. But the principle is the same.
Pick one system and stick with it. What's the worst that could happen?
Last edited:
Dogtrousers
Kilometre nibbler
Not completely.
Haversine is just a lot easier to use when your inputs are GPS co-ordinates. To use simple trigonometry you'd have to convert your co-ordinates to a local projection like the OS grid and then do the trig. And that's harder than just using the haversine formula straight off, in my experience.
Also, sometimes the distance between two co-ordinates does require correcion for curvature. Not in the context of a bike ride GPX, true, but it does happen.
Last edited:
Is that level of accuracy really required in a practical sense
I worked out during a fit of boredom that my trip to work is longer on the way to work than back.
In three cases I go three quarters of the way around roundabouts that mean a further distance. Also on the way home the road lane is closer, again meaning g a shorter distance
I made it some 500 yards!
I worked out during a fit of boredom that my trip to work is longer on the way to work than back.
In three cases I go three quarters of the way around roundabouts that mean a further distance. Also on the way home the road lane is closer, again meaning g a shorter distance
I made it some 500 yards!
jefmcg
Guru
You should go around them more than once,, to clock up some serious mileage.
View: https://www.youtube.com/watch?v=OKGnyh5jIWM
ColinJ
Puzzle game procrastinator!
- Location
- Todmorden - Yorks/Lancs border
Ah, but WHAT sea level?
Here is a photo taken at Knott End on one of my forum rides at low tide. My GPS showed an elevation of -3 metres. The thing is - there were barnacles on that post behind us about 4 metres above the sea level at the time!
Salty seadog
Space Cadet...(3rd Class...)
ColinJ
Puzzle game procrastinator!
- Location
- Todmorden - Yorks/Lancs border
Aravis
Putrid Donut
- Location
- Gloucester
For a few months I've been routinely running 2 Garmins - a 200 and a 520. I have found that over 100 miles they measure about a tenth of a mile apart. Yesterday's experience was completely in line with what I've seen before - 103.7 miles on the 520, 103.8 on the 200. That is a level of consistency that should satisfy anyone.
Out of interest, I've entered the route manually into both RideWithGPS and BikeHike, and the results were 104.1 and 104.0 respectively.
Whatever the theoretical gold standard may be, it seems to me that we have consistency far beyond what was possible with manually calibrated wheel-counters. Also, rides measured by GPS are a matter of record, as opposed to existing only in the rider's memory.
Out of interest, I've entered the route manually into both RideWithGPS and BikeHike, and the results were 104.1 and 104.0 respectively.
Whatever the theoretical gold standard may be, it seems to me that we have consistency far beyond what was possible with manually calibrated wheel-counters. Also, rides measured by GPS are a matter of record, as opposed to existing only in the rider's memory.
