Tuesday, December 3, 2013

GPS Approximations

I recently carried out a USATF certification of the local Turkey Day 5K course. The race promoter wanted the course to have the same start and finish line, so I had to make extra measurements. In fact, I toured the course eleven times to make sure it was as accurate as possible while adjusting for a common start/finish line.

Though I have not kept count, I believe that I have carried out this process more than 150 times dating back to the 1980s. This measurement went smoothly. The temperature remained constant, so six of the eight calibration rides, four before and four after the final three rides, resulted in the same number of clicks. I encountered light traffic, which allowed for tight tangents on the almost constantly arcing course. The three rides on the final course only varied by five clicks. That leaves me quite confident in saying that the race distance over the measured tangents was 3.107 miles with an error of +/- 0.001 miles (+/- 5 ft).

That accuracy and precision was with the Jones Counter. The GPS devices I had with me on those rides resulted in a completely different story. I know this has been chronicled in several articles published in national papers and magazines, but it seems that lot of people have not downloaded their data and then zoomed in on the maps.  That leads them to believe that their prized GPS devices are super accurate, not to be questioned, measuring devices. Sorry. They are not. If they were, the sport federations would allow us to certify courses with them.

Here is a superficial look at how GPS devices work. GPS units receive transmissions from satellites in the GPS network. The receiver and satellite are set up to be in synch, so the receiver (your GPS watch, handheld, auto unit, etc) uses lag time of the satellite signal to determine position. The "Man" lets us have signals that come at time intervals resulting in, AT BEST, an 8-10 meter (25-40 ft) ESTIMATION of where the receiver is at a given time. Please reread that last line. It is important that you have that in mind before you view the satellite map images of my race during that recent 5K.

The shot above seems to indicate that I started on the courthouse steps. I assure you that I was just right of the center of the road. I am impressed with the luck of the two flags being so close together. This is not usually the case.

This second shot proves that I am, in fact, Superman, since I ran straight through two stone and steel walls of the hotel. This bizarre path folks, is the GPS drawing a straight line between two satellite transmissions. The device assumes we travel in straight lines between signal readings, so it often cuts off corners like this. Still, I like the image of me powering through those walls while maintaing pace.

More straight line "curves." This time they indicate that I was taking a shortcut - running through the grass and along the sidewalk. Yep. That's what I did. Shhhh.

The first U-turn. GPS devices have their minor shortcomings, but their largest errors come when there are U-turns or trail switchbacks. The likelihood of the device getting a signal at that pivot point is dependent on speed and, mostly, luck. That means "not very likely." Sure, I saw the cones at the end of the median, but I decided to leap over the median and run straight toward my friend, Jeremy, in a game of chicken. Yeah, that's what I was up to.

The second U-turn. Hey, I got away with it the first time!

The in-race measurement pictured above was 3.06 miles. The corner and U-turn errors seem to easily make up for the .05 mile error.

By the way, the GPS measurements during certification varied between 3.05 and 3.14 miles. That is a difference of 475 ft! (Remember, the Jones Counter had a consistency of about 5 ft)  The route for the longest GPS measurement managed to go beyond one of the U-turns, then had me riding through the river for a quarter of a mile, (Superman. Duh!) and ended beyond the finish line.

The moral to the story is the same one projected in the articles appearing in major publications. Your GPS is a great tool for approximating distances, but it does not measure infallibly. So, please do not use your GPS measurements to tell a race director that their course is too long or too short. You might, instead, ask them if it has been certified with a Jones Counter using the stringent USATF certifying procedure.     ST