You negative split the whole thing!
My GPS normally gives 3.14-3.18 for a measured 5K course. With official measurements, they have to do straight lines, which sometimes may not even be possible given where they will place cones on courses.
Until late 2014, my best time in Rochester was at the Brighton 5K - but I have not run it years. So it is a pretty speedy course, despite doing some of the 'winding through the neighborhood' thing.
Thanks for the insight, Carl. I've often found my GPS agrees amazingly well with certified course distances, but not this time. Although being off 0.03 miles is only 1% -- but still I'm usually closer than that. As you mentioned, this was a fairly "windy" (meaning it winds a lot, not that the wind blows a lot!), so possibly that had something to do with it, either in the way my GPS computed distance, or the way the course was measured during certification, vs the route I actually ran.
I think trailsnail's GPS (different model) said 3.12 miles, which is pretty decent agreement. Maybe she ran a more efficient route than I did!