I would always be inclined to trust a race that has been officially measured with a Jones counter using the most direct racing line rather than GPS on your phone, which is not as accurate, especially if there are buildings in the way, tight bends, etc. If the race has been accurately measured most people's GPS will come out a little long.

There were some cases in the UK where a few races in the same area all turned out to be slightly short; this was a one-off due to the fact that the measuring equipment had been wrongly calibrated.

Pretty much every race I've run has been off. Marathons usually come in .2-.3 too long, and halves about .1-.2. 5ks are typically pretty close, maybe .05 off at the end. I think it's mostly because the course is measured on the tightest tangent lines and most people don't run the exact line.

Mine all come in 'short' on my watch, because its GPS is a bit weird. It partly depends on how frequently your watch or phone updates its satellite signals, and partly on the shape of the course. As an experiment, try running a mile in lane one of a 400m track and see how far your watch thinks it is and where it thinks you've been on a map. Then multiply that level of inaccuracy by whatever your race distance is. And that's not taking into account tree cover, tall buildings, etc. GPS manufacturers don't even claim a precise level of accuracy for their devices; I don't understand why we all (even me) tend to think our watch knows better than the course measurers!

I use my Garmin in training runs. I'll run the same course, the same way, and frequently come up with a different mileage at the end of the run. GPS may be off a little, tangents may be run differently (slightly) on each run, the race measurement may have use different tangents than those you ran, whatever app you were using on line to chart the course knew nothing about tangents, et al.

How much earlier was your GPS to the mileage signs? A couple of yards, a tenth of a mile, a quarter mile or more? It's a shame your GPS crashed, would have been interesting to see what the total mileage for the race was on that.

Does the difference in distance really make that much of a difference?

P.S. I also upload my runs to various different websites, and there is often quite a difference between them, using the exact same GPS data, depending on what their logarithm assumes about how directly I ran between the points.

For me, it depends upon the GPS device and the program. Using my cell phone (which can see up to 22 satellites at any one time but uses between 10-18 for a GPS fix) can give two differing results using two different programs, simultaneously. For example, iCardio (Digifit) tends to read long by between 3 and 6 percent compared to a measured race course distance. However, another program like MapMyFitness or Runtastic Pro tend to read shorter than iCardio and more accurately to a measured course with no more than a 2% "long reading" (1.02 miles instead of 1.00).

Some of this inaccuracy may be caused by trees, buildings and even atmospheric conditions. Cell phones generally do not use WAAS satellite signals to compensate for such effects. And downloading of the csv file (if available) can be revealing in how well your device is actually processing satellite data.

Although I know that most of us do not run the shortest route possible (we zig zag around slow moving obstructions) and such increased distance can add up over the course of a marathon, I start getting suspicious not of the course, but of someone's GPS equipment if they start reporting that the marathon course distance is more than about half a mile longer than the course certification says it is. If you've ever read any of those certification documents, they are pretty specific about the waypoints.

I bought a TomTom GPS/HRM watch this past spring. After comparing the wrist-based HRM data with the chest strap data from another program and satisfying myself that the slight differences in HR readings were not significant for the purposes of HR-based training, I turned my attention to the GPS data. For known distances with plenty of open sky for GPS satellite reception, it seemed very accurate when compared to my phone-based programs.

My experience during race conditions have born this out for a variety of race distances from the 5K up to marathon distance. In my most recent marathon is Venice, Italy, the course measurements were in Km (the mileage was posted in a little box area on the bottom of each sign). As I ran along, I could see that the watch was keeping a very accurate accounting of the course certified distance. At the end of the race, my watch read 26.25 miles rather than 26.22 miles. And this has also given me a very accurate (and realistic) measurement of my pace.

In running/racing conditions, this is now my preferred tracking device and method. Another watch, let alone watch manufacturer, might not prove to be as good. As always, YMMV!

We have a marathon that is a certified Boston Qualifier in March, I had no trouble understanding how the course was designed and arrived at the full and half course designed and mapped by Beast Pacing.

How certified courses are measured and running the tangents

http://www.clevelandmarathon.com/resources/userUploads/CertifiedCoursevsGPSAccuracy.SRV_1338321859.pdf

http://www.dcrainmaker.com/2009/03/racing-line-understanding-how-courses.html

http://www.runninginjuryfree.org/2008/10/run-tangents.html

http://sciencebasedrunning.com/2014/12/the-basics-running-tangents/


http://community.myfitnesspal.com/en/discussion/1145856/course-distance-accuracy

@Stoshew71 sounds like someone who may have certified a course! (As have I) It is super annoying to have people complain about the course being long because their Garmin shows 3.2 miles for the 5k.

No, I have not certified a course before, but I have mad respects for all that do. Just like I have lots of respect for anyone that volunteered their services at any capacity whether it's handing out water to being a race director to being president to a local running club.

.1 mile off is actually in some cases planned purposely (or very close to that) by race coordinators. It's to guarantee that a race is not run too short. Most certified courses are planned to be slightly longer for that reason.

The other thing I try to remember that even if the course is certified, the mile markers generally are not. So yes the entire course should be pretty close but the mile markers are measured out and set up likely with far less precise technology. I consider the mile markers to be just general information.

As a point of reference, large races use professional GPS race mapping tools. However, these tools calculate "shortest possible distance". This means that if you somehow take every turn, bend and straight-away on the same exact plot the GPS tool had calculated, your GPS device will measure the exact same distance. I had tried this out with a few Rock 'n Roll races. Keep in mind though, if you are the type of person that likes avoiding others and dodging and getting around clusters, you may find your distance on your device is longer than the race. Likewise, you may find your final distance on your GPS device is less than the actual race, which happened to me at Pittsburgh Half Marathon a few years ago (I was super impatient and took lots of sidewalks to bypass clusters). And of course I hurt my ankle doing this...lesson learned.

You will find many personal tools like Garmin and Map My Fitness are pretty close to accurate to the professional tools. In addition, you will find that if you dual GPS device on a run that you mapped out on one device, the other device will usually be almost spot on with the mapped device. Sorry for the lengthy response.

Short story long, the certified race distances are almost always correct.

Thanksgiving 5k this year was my 3rd race that regisgtered short on my Garmin. Had a 5k 0.04 short, a HM 0.3 short, and this last one 0.05 short. However, 2 other runners who also track on Strava, showed the 5k this year as right on 5k. I may have run a tiny bit less because we were supposed to stay left side of the road, and I did cut some tangents tighter than that.

I ran a marathon a few years ago with a pacer who is also a certifying official. As someone earlier said, because of the way they have to measure tangents and stay close to the curb it is virtually impossible for you as the runner to follow the path they measured. A properly marked and measured marathon should show ~26.4 on your Garmin. If you get 26.2 the course was short.