http://www.adarq.org/progress-journals-experimental-routines/adarq's-journal/msg125890/#msg125890
gps-timed 400m 
-> 1:04 (-3s)
What's the error on your GPS accuracy-wise?
not sure.. I think it depends on the weather and such. One thing I do is look at the graphs and if they don't have any weird spikes/noticeable issues then it's usually spot on.
for example, here's the 0.25 mi run from last night:
rarely you may seem some weirdness in such a graph.. which would be related to the gps data. I like that one tho, shows me stopped before the run, then boom.. and maintaining speed decently throughout (until the last ~100m).
I've also been using the lap key so, that acts like a stop watch. It said 1:05 for that lap for 0.25 mi .. but it could also be like 0.250 or 0.256 etc (because i'm running on a "stretch" not a track), I only get the precision to two significant digits.
I didn't know about it until recently.. hah.
I'd imagine most are almost to the metre but I've heard from some distance guys that there can be errors up to 20m on a 400m track depending on brand and model.
as for my model, it seems pretty accurate. The thing that has caused errors in the past has been very cloudy weather and stuff like that. Also one freak error where it didn't record a potential 1KM pr.
Someone once told me to expect it to be off when i'm at a track.. i've used it a few times and it's been spot on. So maybe there's more chance of error on a short "loop"? not sure tbh.
I mostly run on sidewalks (and very rarely street roads) so, I usually get large stretches of land.. maybe it works better under those conditions.
edit: oh and the biggest chance for error is right in the beginning.. so if you started a watch right at the start of the run, you may see some weird stuff in the beginning. That's what used to kill me when measuring shorter runs from the "beginning".. once the watch gets going though, seems good.. so then you just use the lap key to find the times of the intervals and bewm.
peace!!
