Jun 6, 2014

Today I want to write about something a little different – GPS devices. These gadgets are a vital part of any attempt to run around the world and I’m generally delighted with what they provide. However, they’re not always accurate.

I’ve been using Garmin devices which, for the most part, perform admirably. And the Garmin Connect web site, which allows me to download all the data in respect to any run and make it publicly available, is a fantastic innovation. But over the past couple of years I’ve come to realise it’s important not to put too much trust in the accuracy of such devices.

When I started my world run I was using a Garmin watch. These are designed for running, but the recharging terminals simply don’t last too long. Mine corroded and was unusable within seven months. Admittedly, I was using it for seven or more hours per day, which is a lot more than just about anyone else.

So I decided switch to a Garmin cycling device. I had to hold this in my hand or put it in my belt pouch, but it had more functions (such as temperature), had a larger and more comprehensive display, and its survival and longevity in adverse conditions was far superior to the watch.

So I merrily went on my way for months with the Garmin cycling device, assuming it was accurately recording my distance and oblivious to anything otherwise. However, from time to time I’d pass sections of highway where the authorities had placed odometer checks. These are accurate kilometre markers meant to enable motorists to precisely check that their odometers and speedometers were working accurately.

But my GPS device was coming up short. Such markers are generally over a distance of 5 km, and my GPS measured only 4.77 km the first time. This is almost 5% less than it should have been. I initially assumed the authorities had been incompetent in the placing of the markers, but when similar losses of distance occurred on other occasions, I started to realise that the problem might actually reside with my device. Any chance after this I’d compare what the GPS was indicating to some supposedly known distance. It was usually short by anything up to 5% or more, though often less than that. For those with good memories, you may recall I mentioned this fact a few times during blogs on the world run.

There wasn’t much I could do about it, and besides, it certainly didn’t invalidate my run since any such inaccuracy meant I was actually running further in total than I was claiming. After the world run finished I sort of forgot about the issue.

Recently I began running with the GPS again. But this time I had enabled a function called Auto Pause. This emits a beep that tells you when the device believes you have stopped, and it stops recording data. As I ran along the road, I’d often hear a beep, look at the screen see the Auto Pause sign on, and notice that the distance was not increasing as I ran. Shortly thereafter, another beep would sound and the Auto Resume sign appeared, after which the distance I’d run would begin to increase again.

Clearly the Garmin cycle computer was not capturing the full distance I  was running. Why was this? I’ve had various discussions with people, and after synthesizing this information, this is my theory (though I could be wrong, so please take it for what it is – a possible reason).

The cycle computer, I believe, is calibrated to operate accurately at speeds at which most bikes travel, that is 25 – 40 km per hour. But I was only running at around 8 km per hour. My understanding is the device records a GPS measurement every half second. During that time, however, I was only covering about 1.5 metres, which is roughly the degree of accuracy of the GPS’s ability to provide positional data. Therefore, even the slightest aberration in its signal might result in it detecting me as not having moved. I know that the accuracy definitely drops off if I’m under awnings, buildings, trees etc., and even if I’m holding it too close to my body.

The reason I’ve decided to mention it today is because I was running this morning in the rain. For this reason I stayed under shop awnings in Randwick for much of the time. After noting that my pace seemed slow, I decided to up my pace. Yet, the next kilometre, according to the GPS device, took almost 8 minutes. I’m experience enough to know when I’m running a lot quicker than an 8 min/km dawdle. I decided to run a well known and accurately measured one kilometre distance nearby at roughly the same pace to compare. It took me 5:25. Clearly, running under the awnings prevented the device from registering my progress for extended periods.

So what does this mean in terms of the world run? It’s impossible to say just how much distance I missed out on. Let’s assume the watch was fairly accurate during the first seven months. That still leaves more than 13 months of running with the cycle device, where I may have missed up to 5% or more of my distance on any given day. I don’t think it was that much on a regular basis, but it was probably an average of no less than 2 %. This is consistent with my experience comparing the device’s distances each day with that of Google Maps, paper maps, and road signs. These alternative measures always seemed to indicate a greater distance than the Garmin was showing. Having covered some 16,000 km during those 13 months, that’s at least 300 km I ran that was not attributed to me (and it may have been more).

So I almost certainly ran further than the official 26,232 km with which I’m credited. But that’s infinitely better than the other way around. If the device had been recording larger distances than I really ran, then my world record would have been invalidated. But not once did the Garmin ever indicate a distance that was further than any other form of measurement had indicated.

The moral of the story is, never put blind faith in anything, including the latest technology. That said, I’ll reiterate, I believe Garmin GPS devices are an amazing technological addition to any world runner’s arsenal. And the cycling devices are probably quite accurate when used at typical cycling speeds.


On This Day


Jun 6, 2012

Distance today = 50.03 km; Total distance = 7368.41 km; Location = Summit, South Dakota (5 km west of) – 45 18.697′ N, 97 06.196′ W; Start time = 0826, Finish time = 1646



Sorry folks, today’s offering is going to have to be short. Lots of other commitments to attend to.

The day involved a TV interview with KDLT, an NBC affiliate, which featured on the South Dakota News tonight. The clip will be up on this web site soon. Also an interview with Terry from Watertown’s People’s Opinion newspaper.

Had a problem with my right leg today – it felt like it wasn’t working, and the left leg had to do all the work – something I’m going to have to work on.

I passed the 45 degree north latitude today, which means I’m now closer to the North Pole than I am to the Equator.

The tracker shows me about 15 km further south than my true position. Also, due to many stoppages today, the watch battery gave out, so there are two Garmin links.

Finally, I’d like to give a huge thanks to the Bittersweet Lodge for putting us up tonight, virtually free of charge. It was a very generous gesture, which we greatly appreciate. The place is great, and I’d honestly say that, even if we had received no reduction – a gem of a place, sitting right in the middle of some lovely rural countryside. It’s very reminiscent of regional France.