GPS Watches: Why and How Marathoners Must Calibrate Them for Happiness

GPS Watches: Why and How Marathoners Must Calibrate Them for Happiness
GPS watches are like speedometers for marathoners’ wrists. But, have you ever gotten a marathon chip-time slower than what your GPS watch was saying? I can explain why and tell you what to do about this.

First, I’ll use my own marathon data to show you how GPS watches can mislead marathoners. Second, I’ll give several reasons why GPS watches overestimate a marathon’s distance. Third, I’ll explain why GPS watches can tell you that you’re faster than you are. Fourth, I’ll show you two approaches for estimating your GPS watch’s distance-measuring inaccuracy. Fifth, I’ll teach you how to calibrate your GPS watch for your upcoming marathon.

GPS watches can mislead marathoners

Here are distances that my GPS watches reported over several of my marathons. I pulled these from the calendar feature on the Garmin Connect website. (Thank you, Garmin!) It had most of my marathons from three Garmin Forerunner watches over the listed years.

The official distance of a marathon is 26.21875 miles. So, two of the marathons in this list got underestimated distances. And, the other twelve marathons got overestimated distances.

The average across the fourteen listed marathons was 26.51286 miles. That’s a marathon-distance overestimation of 0.29411 miles.

Look at the GPS-watch-measured distances, though, in the final three marathons listed. The average across those three was 26.71667 miles. That’s a marathon-distance overestimation of 0.49792 miles.

Why do GPS watches overestimate a marathon’s distance, and why is this bad?

Why GPS watches overestimate a marathon’s distance

A marathon committee hires a road-course certifier to measure the race course. That person uses something other than a GPS unit to certify that the course is 26.21875 miles long. For example, the certifier may use a calibrated bicycle.

Quoting from USATF,

“A race course is defined by the shortest possible route [SPR] that a runner could take and not be disqualified.”

Quoting again from USATF,

“Most runners don’t actually run the SPR, so the distance recorded by their GPS device will usually be longer than the certified length of the course, even though the course was properly measured along the SPR according to USATF rules.”

When I coined “the Mahoney-Meandering-Marathon distance” in 2010, I made an assumption. I assumed that GPS watches measured distance well. So, I attributed the extra distance recorded on my GPS watch as due to my meandering.

But, I now know this:
[Tweet “Distance overestimates by GPS watches need not come only from meandering.”]
Douglas McCormick wrote “Why Every GPS Overestimates Distance Traveled” for IEEE Spectrum. His 2015 article cited researchers at University of Salzburg and two other institutions.

Here’s the first salient paragraph from McCormick’s article about the Salzburg paper:

“Measurement errors have many causes. The paper specifically cites: propagation delay (atmospheric fluctuations affect the speed of the GPS signal); ephemeris error (uncertainty in the precise position of the GPS satellite); satellite clock drift; hardware error (the shortcomings of the terrestrial GPS unit); signal reflections (which can increase the length of the signal path); and unfavorable satellite geometry (available GPS satellites are too low in the sky or too close together or too few, for example).”

And, here’s McCormick’s conclusion about the Salzburg paper:

“Put them together and you have readings that scatter around the true position. The Salzburg researchers found that distances derived from position measurements with randomly distributed errors will, on average, come up longer than the actual separation between two points [emphasis added].”

So, all these measurement factors cause GPS watches to overestimate distance traveled:

  • propagation delay;
  • ephemeris error;
  • satellite clock drift;
  • hardware error;
  • signal reflections;
  • unfavorable satellite geometry.

Why GPS watches can tell you that you’re faster than you are

Pace = duration/distance. Much of the world measures pace in minutes per kilometer. This explains why I wrote Run:Walk Pace Tables in Minutes/Kilometer. But, American marathoners measure pace in minutes per mile. So, as an American, I will use minutes/mile for the rest of this discussion.

A lower minutes/mile value is a faster pace. For example, a pace of 10:00/mile is faster than a pace of 12:00/mile. Put in another way, someone with a 10:00/mile pace will finish a marathon sooner than someone with an 12:00/mile pace.

GPS watches are superb at measuring duration. But, when a GPS watch overestimates distance for a given duration, what happens?

To make the arithmetic simple, let’s look at an extreme example. Suppose that you ran for 12 minutes, and this, of course, is the duration measured by your GPS watch.

  • If you covered in reality one mile, then your true pace was 12:00/mile.
  • If your GPS watch estimated that you covered 1.2 miles, then it would report your pace as 10:00/mile. (12 minutes/1.2 miles = 10:00/mile.)

In other words, …
[Tweet “When your GPS watch overestimates distance, it tells you that you’re faster than you are.”]

How to estimate your GPS watch’s distance-measuring inaccuracy

My list showed how variable GPS watches can be in measuring the distance covered in a marathon. But, you should create your own list.

The Simple Approach

Suppose that you have several marathon distance measurements from your GPS watch. You could sum those distances and divide by the total number of marathons. This would give you the average of your GPS watch’s overestimated distances.

The More Sophisticated Approach

If you have enough data, you can take a more sophisticated approach to getting the distance.

  • Is it the same marathon as your upcoming marathon? Marathons differ in topography. For example, Texas Marathon follows a wooded trail. In contrast, Chevron Houston Marathon follows an urban and suburban course. The first gets satellite-signal reflections from nearby lakes. The second gets those reflections from tall glass buildings in parts of the course. Use only previous years for the same marathon to get the same “signal reflections” error.
  • Is it the same course for the same marathon? Suppose that you have completed the same marathon for several years. But, suppose that the marathon committee changed the course at some point. Then you could use only the data from marathons that match the upcoming marathon’s course. For example, look at the Chevron Houston Marathon. Its course for 2014 and later differs from its course for 2013 and earlier.
  • Is it the same GPS watch brand? One could imagine that different watch brands use different algorithms to estimate distance. After all, manufacturers compete on accuracy as well as looks and functionality. So, you could use only the data from marathons that match the brand of watch that you plan to use in your upcoming race.
  • Is it the same GPS watch? You could use only the data from marathons that match the watch that you plan to use in your upcoming marathon. For example, my list includes distance measurements from three different Garmin watches. Each one has been great in its own right. On average, though, I’d hope that a newer-model GPS watch would be more accurate than an older one.
  • Is it the same degree of meandering? As discussed, meandering may not contribute the most to distance overestimation. But, you could use only the years that match the degree of meandering that you expect in your upcoming race. For example, suppose that you started toward the back of a chute in the early years of a marathon. And, suppose that you have been starting toward the front of a chute in recent years. This could affect how much you must meander around other marathoners in a race. As another example, suppose that you are more assertive now than you were in your early years. You may tend to run the tangents more now, which would lower your meandering. You could decide to include only data from marathons with assertive runs of tangents.

How to calibrate your GPS watch for your upcoming marathon

Step 1. Decide what “overestimated distance” you will use. No matter which approach you use or how sophisticated you choose to be, you need a distance. That is,
[Tweet “What is the distance that you expect your GPS watch to show at the end of your upcoming marathon?”]
Let’s use “overestimated distance” to refer to this distance that you expect. For our seven-steps example, let’s use the Mahoney-Meandering-Marathon distance of 26.65 miles.

Step 2. Choose a desired chip-time for your upcoming marathon. Continuing our example, let’s say that you want to finish in five hours.

Step 3. Look up or compute your “official” pace for your desired chip-time.
[Tweet “Your ‘official’ pace equals your desired marathon chip-time divided by 26.21875 miles.”]
Continuing our example, the official pace for a five-hour marathon is about 11:27/mile.

Step 4. Convert your official pace from minutes and seconds/mile to seconds/mile. For example, 11:27/mile = 687 seconds/mile.

Step 5. Compute the overestimated-distance pace in seconds/mile. To do this, first multiply your seconds/mile official pace by the official distance. Then divide the result by the overestimated distance. Continuing our example, our seconds/mile official pace = 687 seconds/mile. Multiplying this by 26.21875 miles and dividing the result by 26.65 miles gives us 675.88 seconds/mile.

Step 6. Convert the overestimated-distance pace from seconds/mile to minutes and seconds/mile. For example, 675.88 seconds/mile = 11:15.88/mile.

Step 7. Round the overestimated-distance pace down to the nearest whole second. For example, 11:15.88/mile rounded down = 11:15/mile.

So, if you want to finish your upcoming marathon in the desired chip-time from step 2, then…
[Tweet “Your GPS watch should display an average pace equal to your overestimated-distance pace.”]
Summarizing our example for a five-hour marathon, …

  • Your official pace must be 11:27/mile;
  • Your overestimated-distance pace must be 11:15/mile.

A 12-seconds/mile difference (11:27 vs. 11:15) in pace may not sound like a lot, but it is. If you kept seeing an 11:27/mile average pace on your GPS watch in this example, then you would not finish in five hours!

You may be wondering…
[Tweet “Is an overestimated-distance pace always 12 seconds/mile faster than an official pace?”]
No.

That 12-seconds/mile difference comes from an overestimated distance of 26.65 miles. And, it comes from a desired chip-time of five hours.
[Tweet “You must complete my seven-step process to compute your GPS watch’s overestimated-distance pace.”]
Let’s look at three examples. With 26.65 miles again as your GPS watch’s overestimated distance, …

  • A 6:00/mile official pace equals a 5:54/mile overestimated-distance pace.
  • A 10:00/mile official pace equals a 9:50/mile overestimated-distance pace.
  • A 14:00/mile official pace equals a 13:46/mile overestimated-distance pace.

The first example involves a 6-seconds/mile difference. The second example involves a 10-seconds/mile difference. The third example involves a 14-seconds/mile difference. And, all these examples would change with a different overestimated distance.

If you follow my seven-step process, then you will know what your GPS watch should display as average pace. Suppose, though, that you don’t have any marathon data for step 1. Then you could use 26.65 miles as your overestimated distance. Even better, my book Run:Walk Pace Tables in Minutes/Mile would let you skip the other steps, too. The reason that you can skip all steps? The book tells you Mahoney-Meandering-Marathon (“MMM”) chip-times for many overestimated-distance paces. For example, you need an overestimated-distance pace of 9:00/mile to finish in 3:59:51. Plus, because this book targets run:walk athletes, there are MMM tables for run:walk ratios from 1:1 through 10:1.

Interested? Click here:

YES! I want to get Mahoney-Meandering-Marathon chip-time estimates!

Image Credit: Pixabay

Posted by

Category: Run:Walk Method