Lies, Damn Lies, and GPS Data

This is also a subtitle of a section of an article about measurement, but I think it applies to what I'm about to say. I had considered using "Sex, Lies and GPS Data, without the Sex" as the subject, but I like the Mark Twain thing better.

I got my Forerunner 305 toy a few days ago, and now I've had a couple runs with it. I'm still in the process of figuring things out, but I've noticed a few things and have a few questions right off the bat.

For one thing, based on doing some of my standard courses with the device, everything is shorter than I thought. Even the courses that I measured on gmaps pedometer. Even the ones where I drove the course. This may be because I cut tangents when I run. Or it's because I'm losing the signal under trees and the device thinks I'm cutting tangents when I'm not. Regardless, I find it mildly upsetting to learn that my 6-mile course is really 5.85 miles, and that my 11-mile course is really 10.87 miles.

Here's something weird: I've checked the distance from my house to the end of the street several times. I've come up with .28, .30 and .31 miles. I can imagine that it all depends on EXACTLY where I started and where I crossed the street, etc. But when I measured it walking, I came up with .34 miles. Everything else seemed longer by walking than running as well. These are just preliminary findings, but has anyones else experienced this?

Another thought: This may be a silly question, but does arm movement have an effect on distance?

One other initial finding: after determining that at least some of my courses are shorter than I thought, I figured that running at the track would come up with distances much shorter than I've logged. This is for a couple of reasons: 1) running four laps in the inside lane is indeed less than a mile, and 2) with all the turning, the forerunner may lose some distance and determine tangents. But here, the distance measured was *long*. The first 4 laps showed up as 1.04 miles. Subsequent miles were also long. What's this all about?

The main thing I'm doing here initially though, is gathering data. This early analysis (especially when I map it out) is interesting, but I can do more later. This applies to the HRM data too. That's another story.

ORN: Just my 3-mile course today. This one actually did measure out very close to what I thought. This was my most decent run of the week, since the marathon. Now I've got a 4-mile race tomorrow.