Got this as part of an email from the Good Run Guide today. Given so many comments of "my garmain said..." etc I thought it was worth sharing.
Measuring RoutesHow far did you run today? Unfortunately this question is not as easy to answer as it might seem. However we all measure our runs, it is virtually impossible get a precise measurement. All measurement systems have an inherent level of inaccuracy and none can measure the exact line that a runner took around a route. So, how accurate can we expect to be?
We looked at several popular methods for measuring routes and compared general levels of accuracy.
1.Jones Counter
This is the benchmark for accurate measurement of running routes and is the method used to certify the accuracy of many UK and International races. A Jones Counter is a small device attached to a bicycle wheel that measures distance by counting the revolutions, or partial revolutions, of the wheel. It is not something that can be used by an everyday runner but achieves accuracy levels of +/- 0.1%. Obviously if you deviate from the measured line, the distance you run will differ from the route measurement.
2.Online Route Mapping
We tested the accuracy of the Good Run Guide Route Measurer by measuring four laps of a 400m athletics track using the satellite map view (see picture above) and repeated the exercise 20 times. The measurements ranged from 1598.344m to 1603.518m, all within 0.22% of the correct distance. The vast majority were within 0.1% of the correct distance, although accuracy levels do depend a lot on plotting points in the right place!
3.GPS Devices
We repeated the same athletics track test with two popular running GPS devices. The test was carried out on two different open air athletics tracks, each with a clear view of the sky giving a good GPS signal. Two runners recorded 20 separate measurements of a 1600m route by running four 400m laps of the track. The measurements ranged from 1571.693m to 1734.660m. The maximum innaccuracy was 8.4% but the majority of measurements were within 4% of the correct distance. These tests were conducted with good GPS reception, so in areas where GPS signals are obscured or reflected from buildings, additional inaccuracies could be introduced if satellite signals are delayed or prevented from reaching the device.
4.Pedometers
Pedometers estimate how far you’ve run by counting your steps and using an average stride length. While they can give a guide to how far you are running, high levels of inaccuracy are not uncommon; Ghent University carried out a study on 1000 pedometers and found that most were inaccurate. Three in four were over 10% inaccurate, one in three were more than 50% inaccurate.
5.Footpods
Footpods are small devices worn in or around the running shoe that measure distance by detecting the motion of your foot. They are also not dependent on satellite signals so can be used anywhere. We haven't tested these oursevles but manufacturers claim accuracy levels of around 1% to 2% after calibration.
Summary
No method of measuring routes is perfect. As a guide, the following table shows indicative levels of accuracy using different methods. However, as it is almost impossible to measure the exact route you actually took, we recommend you always allow for a potential error factor of at least 0.5% regardless of which method you use.
Comments
MethodAccuracy Level Guide *
Jones Counter+/- 0.1%
Satellite Map Measurement+/- 0.25% [1]
Foot Pod +/- 2% [2]GPS Device+/- 4% [3]
Pedometer+/- 10% [4]
* Indicative only - different levels of accuracy may be achieved in different circumstances
A number of comments were made on the Llanelli half today that Garmins were measuring 13.3ish, with some people therefore assuming that they'd run further and that their times were even better than those published! This is only a 1.5% inaccuracy, so well within the average inaccuracy of a GPS Device. I had often wondered at why runners assumed their garmins were perfect whilst other measuring systems weren't, as though they'd never driven with a GPS system in the car!
Very interesting. I do hope to see a little less of the "My garmin said it was this...", and if not, I'll point people this way!
In my experience Garmins are very consistent - I can't say accurate - but certainly using one on my bike it measures certain points (like a local cafe etc) as an indentical distance every time.
I don't know how they work - for example do they take a continuous reading or repeated sample points and plot a route from that smoothing the line between the points ? If it's the latter I suppose that might explain why running in circles on a track might not see them at their most accurate.
I use GPS when I do training runs (I have an app for my iPhone), sometimes I just run and let it measure, and sometimes I'll plot myself a set route to follow. I'm always prepared for it to be a bit out but for my purposes I don't find that it really matters.
If I turn up on race day and there's an extra couple of metres on top of what I'm used to to running, I'd like to think I could cope
Inaccuracies in distance measurement are probably just one of the things you have to put up with when training outside (like the wind and rain, and potholes!), but I would rather have that than have to run on a treadmill!
Running round a track to test GPS devices is a schoolboy error. GPS measures in short straight lines, so any course with a lot of curves will not be measured accurately,
Anyone who uses a GPS watch and later looks at their downloaded route on a map knows that there is often a cutting off of corners, because of the way that the device takes frequent 'snapshots' of your position. If a snapshot happens just before a corner, then just after a corner, the GPS can't possibly know that you went around the corner. It just assumes you went direct from point A to point B.
Using a GPS watch on a track will be hopelessly inaccurate for this reason. Depicted on a map, the route around the bends will look like a series of short straight lines, and not a nice curved line.
GPS is pretty accurate on the sort of courses that most runners normally use.
I used to use the Nike + footpod, even when I ran the same route I would always get different distances. At times when it was really bad I would lose or gain half a km. I used to to give me a general idea of distance but always took its result with a pinch of salt.
After being caught out in a nasty rain storm, the Nike + (and the ipod it was attached to), stopped working so I switched over to Mapmyrun.com. Once again the accuracy comes down to how well I can plot my route out but until I decide whether or not to buy a Garmin, it's not for a free tool!
mapmyrun will work very well if you are running on roads. I have often used it, usually to measure a route in advance. Off-road, obviously, it's not much good unless you have a very clear idea of the path (e.g. a canal towpath)..
GPS won't necessarily be more accurate but it does have the convenience factor. I'm no Garmin evangelist but I've been using GPS for about 5 years now, and on balance, would say it works very well -- but not on a 400 metre track!
The earlier comment about the half marathon coming out at 13.3 is about right. No runner ever runs the exact line of a course, and there are loads of times when you veer off it e.g. going round corners, stopping at drinks stations etc. The GPS reading will almost always be slightly over.
Well said. So should we trust any of their other conclusions either?
I tend to stick to the roads for 95% of my running but yes can understand why mapmyrun would be very hard to use if going off road.
About 3 1/2 -4 years ago, I had this massive GPS device. Big blue triangle which you wore on your upper arm, it might have even been from Garmin. Not a bad bit of kit actually for those technological times but didn't work well if it was a cloudy day. Living in England, it did become a factor!
Jools - It's not a myth but it's not quite correct. AIMS suggest measuring 0.1% over (i.e. 10010m for a 10K) according to the course measurement guide.
RC - You are on the right lines. Measuring a track with a Forerunner or similar is a waste of time. The actual reason is the frequency of measurement. If you measure once every 1-2 seconds (as these devices do) then the length of the track will depend on when you crossed the start/finish line. Other GPS devices (usually using differential GPS) measure far more frequently (I've used devices measuring up to 20Hz) and use intelligent interpolation but are generally used for other purposes. They could however be used to measure the accuracy of the track to less than 1cm if required. GPS is not inherently inaccurate per se, you just need to use the right tool for the job.
Cougie - Differences between a GPS measured route and a map are most probably due to co-ordinate transformation differences between the GPS and map algorithms.Don't worry about it, it's probably not relevant compared to other measurement errors.
Edit: X-Post with Mr.W!
Is there any easy way to "borrow" a Jones counter? I'd quite like to measure one of my favourite routes accurately, and compare that to the map-based and garmin-based measurements I have.
(The route in question is a circuit of Richmond Park, which is quite "wiggly", and hence perhaps prone to the same rounding errors as doing circuits of a track. There are also trees.)
I think the 405 is more accurate (at least mine has been pretty much bang on in the measured races I've done), hence the need for less positional fixes.
Might be wrong, could do with finding the info on where I read it.........
Mike, you need to find a friendly course measurer and borrow theirs. However, it isn't as simple as bolting it onto your bike and riding round the course. The correct procedure for an accurate measurement involves measuring out a calibration course of several hundred meters (I can't remember exactly how far) using a steel tape or ruler then riding the course to check the calibration. Then the full course is ridden several times before the calibration is checked again on the calibration course.
If anyone is interested in doing it then there's a website - http://www.coursemeasurement.org.uk/measurers/index.htm
MF - if you want to 'borrow one without borrowing one' you can do laps of the parkrun course as its 5k (measured with a jones counter) - look into parkrun freedom - I think its gone nationwide now.
I think the point is that for training purposes for most people mapmyrun is accurate enough - I use the satellite mode for off road, but it can be a bit of guesswork.
Of course you dont get pbs from training runs, so here's where an accurate course comes in handy to compare with your last run on an accurately measured course. Its one of a number of reasons that regular racing can be beneficial (provided its not all long distance) to training because it allows you to peg your progress against your stated aims.
But I really hate it when people have set over optimistic targets and then bang on about the course being long - get over it and get back to training!