Got some first hand experience in the reliability of Strava rankings this weekend. Might be due to GPS variance but there may be more to it. Data collection and analysis may have something to do with it too. Either way, myself and others know the truth as of this weekends results. I give you exhibit A, a segment known as Riverview KOM. This is the result of my riding partners upload. Notice anything strange?
How did Juan average 21.7 MPH when his max speed on the segment was 19 MPH? Even better… how did these results rank him at 26th when all present on the ride know I chased a slower rider in a forward group to the top of this hill segment, in a brazen attempt to see if I could summit first and then I literally waited at the top of the hill for Juan to get to the top. I stopped and waited. And it wasn’t even recorded as my personal fasted time on that segment by my GPS. So is my Garmin Edge 500 to blame? Or is Juan’s iPhone 4s the culprit? The next questionable segment sheds some light on the blame.
Here is another very odd result of this ride that does in fact decide a KOM. I’m happy, almost proud, Juan got this KOM because it was his first ever but…. there’s another problem. I pulled Juan on this segment. The entire segment and in fact the entire road the segment is part of. His first comment to me upon seeing this KOM was “but you pulled me”. This was so close that I can only attribute the 10th of a second difference to GPS variance because I was indeed in front, and occasionally creating separation up to 2-3 bike lengths that Juan would have to make up. Further segment analysis proves data collection is the likely offender.
Here are the results of this KOM segment shown from Juan’s Strava profile and my own. Notice Juan’s top speed, 44.3 MPH. I assure you we never did 44 MPH on this segment. I’m confident because my bike was in front and my GPS reflects the more accurate 27.5 MPH top speed leaving me to believe it was Juan’s iPhone 4s skewing the results, giving Juan the KOM.
Before jealousy and animosity are presumed please understand while I care if the rankings are correct, I could not care less if they result in accolades for myself. Like I said I’m happier Juan got his first KOM regardless of how it occurred technically than if I got another. However I will no longer put a lot of faith in top Strava rankings separated by a second or two since I now have personal experience and first hand evidence there is some level of GPS reliability or data analysis skewing results.
Hey, thanks for commenting on my blog, I thought I’d come over and check out your post. For me, the take home lesson here is about the validity of very short segments. There’s always going to be a margin of error, so the shorter the segment, the greater the inaccuracies. I’d go so far as to suggest that below a certain distance/duration, certain segments should be dismissed as irrelevant.