I got so tired on the daily, of explaining to people that when they run a certified course and the Garmin says something else, they should trust the course and not the Garmin. But just about everyone swears by the Garmin no matter what you say. They say, "oh but the software is newer..." and so on. Following is a post from a guy who made such a cogent explanation here, that I saved it, and I send it to anyone who gives me that Garmin nonsense:
"In my boring real life I have a small land surveying practice and have used GPS technology on a regular basis for the last 12 years or so. With GPS, the effective way to measure accurate distances is to use multiple GPS receivers simultaneously and then process the resulting vectors using some high dollar software or even higher dollar "real time" equipment. (The other way is to use one receiver and let it collect data for 23 hours and 56 minutes at each location.)
Obviously this doesn't work for running. I don't have a Garmin and have only seen one at a distance, but I would find it hard to believe that something in the price range of $300 that you wear on your wrist would have the capability to process more than 8 satellites and probably only the P1 signal of those 8. (For the record the US has 24 GPS satellites in 6 orbits of 4 satellites spaced roughly 90° apart. The USSR's version is called GLONASS, while the European Union has navigational satellites as well. I don't remember what they're called.) These US satellites are not fixed, like communication and TV, and their positions change constantly.
One thing I've never seen in a running forum discussion of GPS is "satellite geometry". In reality, in any given 24 hour period, there may be 4 or 5 hours where maximum satellites are spread out over the entire sky, instead of either clustered to one part of the sky or minimum satellites are visible. When tree canopy or urban canyons are thrown into the mix the signal degrades that much more. Surveyors such as myself when planning GPS work, use an almanac or ephemeris, which we download from GPS receivers, to find the optimum time to work. I looked on google, hoping to find a link I could share here, but was unsuccessful.
Basically, if you ever have the chance to use a GPS almanac or ephemeris, you'd have to enter the appx latitude, longitude, and elevation of your location and you would be told the optimum time of day or night for GPS usage. If the data is provided in a graphic form, you want PDOP (positional dilution of precision) to be as low as possible, definitely under 4, and you want the maximum number of satellites available. Right now in southern Oregon, the optimum time is roughly from noon to 3PM local.
Even with the sophisticated equipment we use, I regularly find unadjusted positional data from any single receiver to be at least 50 to 150 feet off, and this is with a 15 minute observation. When you multiply this by 2 (starting and finishing), you can see it's easy to be off 100 to 300 feet when attempting to measure the distance you just ran, be it one mile or 26. But hey, before the US government turned off selective availability in 2000, that 50 to 150 feet was 300 to 500 feet.
Bottom line, it's a neat tool, but it has limits. For myself, I run the same routes and don't much care whether my 10 miler is 9.8 or 10.2 and when I'm in a strange area, I run for time."
BTW, I have a Garmin and use it primarily for running routes that are unfamiliar to me, such as when I travel. The rest of the time, my Timex tri watch is just fine. I can't imagine wearing a Garmin in a race.
Spareribs
