Author Topic: Mileage discrepancies  (Read 2350 times)

RogerT

  • Playing with a big steamy thing
Mileage discrepancies
« on: July 14, 2008, 12:36:31 pm »
I have always noticed a small difference in mileage recorded on my Legend HC and the Cateye computer.  ( about 5 kms in 100 ).  I always put it down to the cateye not being that accurate.

But, on An Audax yesterday the mileages shown on the cateye EXACTLY corresponded with the route notes whereas as the Garmin was about 5% in advance and at the end of a 103 KMS ride showed 109Kms.

I always though that GPS was VERY accurate.

Any ideas/thoughts ?


Re: Mileage discrepancies
« Reply #1 on: July 14, 2008, 12:46:39 pm »
I have always noticed a small difference in mileage recorded on my Legend HC and the Cateye computer.  ( about 5 kms in 100 ).  I always put it down to the cateye not being that accurate.

But, on An Audax yesterday the mileages shown on the cateye EXACTLY corresponded with the route notes whereas as the Garmin was about 5% in advance and at the end of a 103 KMS ride showed 109Kms.

I always though that GPS was VERY accurate.

Any ideas/thoughts ?

The Audax organiser may use a similarly setup cycle computer. It's no guarantee that either computers are accurate.

GPS aren't perfectly accurate for distance calculations. Usual point accuracy is about as good as 5m. Use too many points and the inter-point distances may include too much error. Use too few points and you lose distance where you didn't travel between those two points in a dead straight line. There's clever stuff inside to average things out and keep it sensible, but since there is some possible error in the position data there's no way to get an accurate distance travelled figure.

Even working out distances from maps isn't accurate, unless it's pan flat. A 1km stretch of 5% gradient will still look like 1km on a map, but will actually be 1001.25m along the road.
"Yes please" said Squirrel "biscuits are our favourite things."

Really Ancien

Re: Mileage discrepancies
« Reply #2 on: July 14, 2008, 01:10:15 pm »
I have always noticed a small difference in mileage recorded on my Legend HC and the Cateye computer.  ( about 5 kms in 100 ).  I always put it down to the cateye not being that accurate.

But, on An Audax yesterday the mileages shown on the cateye EXACTLY corresponded with the route notes whereas as the Garmin was about 5% in advance and at the end of a 103 KMS ride showed 109Kms.

I always though that GPS was VERY accurate.

Any ideas/thoughts ?

Try talking to your companions on the road, or failing that, read up on the area you are passing through and study the features of the land. This will enliven your ride and prevent you focussing on minor details.

Damon.




andygates

  • Peroxide Viking
Re: Mileage discrepancies
« Reply #3 on: July 14, 2008, 01:14:28 pm »
The Audax organiser may use a similarly setup cycle computer. It's no guarantee that either computers are accurate.

+1 to that.

Your reported mileage tallies with the reported mileage of the bloke who wrote the sheet.

That doesn't necessarily have anything to do with the real world :D
It takes blood and guts to be this cool but I'm still just a cliché.
OpenStreetMap UK & IRL Streetmap & Topo: ravenfamily.org/andyg/maps updates weekly.

rr

Re: Mileage discrepancies
« Reply #4 on: July 14, 2008, 01:30:45 pm »
We did computer calibration, vs GPS and maps elsewhere in great detail the concludsion was that we trust a well calibrated computer more than anything else

Edited to fix link

RogerT

  • Playing with a big steamy thing
Re: Mileage discrepancies
« Reply #5 on: July 14, 2008, 04:00:26 pm »
I have always noticed a small difference in mileage recorded on my Legend HC and the Cateye computer.  ( about 5 kms in 100 ).  I always put it down to the cateye not being that accurate.

But, on An Audax yesterday the mileages shown on the cateye EXACTLY corresponded with the route notes whereas as the Garmin was about 5% in advance and at the end of a 103 KMS ride showed 109Kms.

I always though that GPS was VERY accurate.

Any ideas/thoughts ?

Try talking to your companions on the road, or failing that, read up on the area you are passing through and study the features of the land. This will enliven your ride and prevent you focussing on minor details.

Damon.





 ;D ;D

I only asked out of curiosity, I am not that anal  ( i think )

I was more interested in the WW2 Piper Cub that was being prepared for flight on the Tibenham Grass strip.  Brought back many memories as I owned a small share in it many years ago.  No GPS in those days, just Mk1 Eyeball.

frankly frankie

  • I kid you not
    • Fuchsiaphile
Re: Mileage discrepancies
« Reply #6 on: July 14, 2008, 05:25:36 pm »
The HC range has a known issue with logging distance covered at very slow speeds - walking pace or slower - but I don't imagine that you were going that slowly, and in any case, this would result in the GPS displaying a lesser distance, not greater as yours did.
(The current latest 'software update' available from the Garmin website is supposed to improve this problem, but reputedly introduces other issues, so personally I haven't applied it to my HCx.)
"This is a complex subject, with a need for more than one highlighter pen."

Re: Mileage discrepancies
« Reply #7 on: July 14, 2008, 05:27:51 pm »
(The current latest 'software update' available from the Garmin website is supposed to improve this problem, but reputedly introduces other issues

What are these, can you linky me?


Re: Mileage discrepancies
« Reply #8 on: July 14, 2008, 08:56:23 pm »
What I'd call quite serious ones, giving rise to positions that can be 2 or 300 metres out. These also arise with the Colorado, and I'd assume therefore with the Oregon as well.

HCx thread on groundspeak
Colorado thread on groundspeak

Maladict

Re: Mileage discrepancies
« Reply #9 on: July 14, 2008, 09:00:13 pm »

Even working out distances from maps isn't accurate, unless it's pan flat. A 1km stretch of 5% gradient will still look like 1km on a map, but will actually be 1001.25m along the road.

Tracklogs has altitude data and has an option to correct for this.

Of course on some types of route it overestimates the climb so it gets that wrong too.  ::-)

Re: Mileage discrepancies
« Reply #10 on: July 14, 2008, 09:03:54 pm »
What I'd call quite serious ones, giving rise to positions that can be 2 or 300 metres out. These also arise with the Colorado, and I'd assume therefore with the Oregon as well.

HCx thread on groundspeak
Colorado thread on groundspeak

Thanks andrew_s.

I have installed the available updates and I have never noticed such discrepancies; maybe they have fixed it.

Re: Mileage discrepancies
« Reply #11 on: July 14, 2008, 09:35:58 pm »
A well-calibrated cycle computer is far more accurate than GPS in terms of distance travelled.

GPS is accurate (to specified limits) in *position* only. Unfortunately, the position wanders a bit from fix to fix - which would appear as spurious distance travelled in a simplistic unit.

This 'wobble' has to be taken into account in software, and there's no exact solution - so GPS units try to filter out the wobble, but inevitably the distance calculation suffers. Ideally, you use a GPS unit with software tuned to the kind of speeds and distances you will be doing (walking units, in theory, use different tuning to cycling units).

Most GPS units seem to use some kind of "keep track of distance variations under X metres, as soon as over X metres add f(X) to total". As a result, with GPS your mileage will vary and depends on the kind of riding you do (so don't expect any kind of consistency with others over the same course).

Addendum: And if you think GPS is bad at tracking distances, check out accurate their 'total ascent' figures are! (not for nothing do any height-specific ones come with an old-fashioned barometer built in).


frankly frankie

  • I kid you not
    • Fuchsiaphile
Re: Mileage discrepancies
« Reply #12 on: July 15, 2008, 10:44:46 am »
What I'd call quite serious ones, ...
HCx thread on groundspeak
Colorado thread on groundspeak
have installed the available updates and I have never noticed such discrepancies ...

All these fora seem to breed paranoia.  2 or 3 people post their bad experiences and the silent majority remain ... silent.  Then someone subjects the unit in question to close scrutiny (as opposed to real-world usage) and lo - there is a measurable 'issue'.
(And the people on Groundspeak are, mainly, concerned with getting a static 'fix' - where we are more interested in following a route - as hazeii says, different GPS models are optimised in different ways.)

FWIW I now have a Legend C, Cx and, as of 3 weeks ago, a HCx (Sheila and I use one each and we carry one as a spare between us) and I don't see any significant difference between them in use, they all seem to me to work very well (though that doesn't stop me having a wish list as long as my arm). 
The x models have a slightly brighter/contrastier screen and of course the memory card (which is a big plus) - the HCx model gets a fix faster than the others if you time them side by side, but its only a matter of 3 or 4 seconds, not significant in practice. 
They all perform well on the bike under tree cover - again the H is measurably better but there is no practical difference at cycling speeds.   They all have battery runtime which is 'good enough' ie 24h+ though I believe the older C model is the best here.
Both the older ones have been known to switch off due to road vibration - but not recently, with a bit more attention to the way they are mounted on the bars - I don't expect the new one to be any different.
"This is a complex subject, with a need for more than one highlighter pen."