Recorded start to finish line too and only just managed to edge you out. My GPS has 10.03k. Under tree cover I am less confident with its readings than my old one but I have been running lots of loops of a set route and think it is pretty reliable on roads. Although the old was was better still in repeatable accuracy, think I had a very reliable 205.
I also have good confidence in my low-end GPS (FR10) on open roads, as evidenced (among other things) by every mile split agreeing almost perfectly (within a few meters) with the mile markers on the course. Thus I was a little surprised to have it say 6.23 miles at the finish line, which is almost exactly the 10.03 km that you measured. That's not off by much, only 30 meters, but when everything had been right on up to that point, it seemed off more than it should have been.
Ok, Since you both have data here,I thought I'd look at mine... I have a split at the 6 mile mark, but didn't cut off when crossing the line. But - the data shows exactly where I slowed and curved after crossing the line. That is a .225 miles...
It would not surprise me if the finish has wandered a few meters over the years.
Another data point is from trailsnail's GPS. Her log says she ran 6.25 miles, but that is from where she actually started, not the start line. But she hit lap when she crossed the starting line, then stop when she hit the finish line. Her second "lap" (which is the start to finish distance) was 6.23 miles.
So it sounds like we have the GPS's of three people agreeing amazingly well, and saying that the course was a bit long! I have to believe that it really was in fact long.
Not a big deal, but when the course is advertised as "certified" it really shouldn't be off. I think I recall that a certified course is expected to be accurate to within 0.01%, which would be 10 meters in a 10K.
(I guess another possibility is that none of us took the absolute shortest route, around corners, etc. But trailsnail and I were pretty careful to always be on the inside at the few corners that there were on the course, and I don't think we did any significant weaving back and forth in the road which would have added some distance.)
Correction, we have four GPS's, not three, all agreeing .... (Delyn, trailsnail, ccsteve, and me)
Oh - yes, actual vs optimal. We started on the left side of the road, and certainly did some wandering... I went to the edge for water, and we ran mostly in the center of the road for the first half. we weren't close to the side of the road until after the last water.
If one were to start on the right, and stay on that shoulder for the race, I might reasonably think there's some distance to cut out.
And - if anything, they are ok with a few meters of overage in a course.
I had heard once (can't remember if this is right or not) that a certified course was to be measured at 0.01% longer than the target distance, which again would be 10 meters in a 10K. That is to account for the presumed maximum error of 0.01%, so that the course will be a minimum of 1X the target distance, and a maximum of 1.002X the target distance.