Help with Strava ;-)

[cxagent]

So I finally relented and started using Strava a little bit. And found something I call BS on. Check out the Strava times in the attached screen shot.

Does anybody really think somebody rode, ran or even flew down rock road at the Spyglass entrance of BCGB in 4 seconds? At over 50 miles an hour? 

I have to think that somebody is "editing" the GPS file. I can see how someone might drive their car on some of the road segments and get this kind of time. But not on this rock utility road. 

[Kyle]

This often happens with really short segments.  When the segment is only 0.06 miles long the effect of a single GPS glitch can be relatively huge and the times and calculated speeds get weird.

[fontarin]

Yeah, it's buggy sometimes on really short segments like this depending on how often the GPS samples.  See the same thing with the short climb up the dam at Walnut. 

[Anita Handle]

If you click on some of those faster times, you can see that the portion of the GPS route that is being given credit for this Strava segment is extremely short or actually runs perpendicular to the segment. Like, they actually rode past this climb on the Greenbelt or on the street and Strava said it was close enough. I recall vividly my best effort up that and I'm not sure I could come anywhere near it today! 

Here's an example of a "7s" effort:

 

[Anita Handle]

Here is something that is realistic and towards the higher end of what a human can do, Tristan Uhl's. Even his GPS track looks a little off from the official track. Still, 18s was his time. I could see someone doing 13 or 14s. Not 5 or 7 though. My fastest was 26s and I always assume that these fast guys could double my pace on something like this.

 

[cxagent]

Wow that "7 second" track is not even close. It is roughly perpendicular to the actual climb. But apparently one point on the climb is all Strava needs to "know" what segment is what. I still question if anybody was doing 50+ mph on the road to get those Strava times.

I actually climbed up the new switchback climb instead of the rock utility road. Since they are roughly parallel and near each other, I could understand Strava thinking they were the same.

Edited February 19, 2019 by cxagent

[cxagent]

On Tristan's track you can see where he looped back to get a good run at climbing the hill.

Not sure why it shows him crossing Spyglass at the top of the hill. Stopped for a Taco maybe? It looks like all of this GPS points had a lot of error in them. That track would be hard to ride because of all the trees and brush in the way. Not to mention the gully to the north of the utility road.

[fontarin]

I find more value in it when I'm looking at improving times on whole tracks, like a full named section of trail.   I pretty much ignore PRs/times these days unless I'm comparing actual race courses or big efforts like hills, and even then it's usually 'oh look, I climbed this faster 5 years ago when I was in shape and weighed 30 lbs less, who would have thought?'.   I just use it to keep track of miles and where I ride.   I really like the individual heatmap feature that shows me where I've ridden and I can compare years.

[Anita Handle]

yes, people invest some serious emotional wattage in these short sections yet Strava is doing its best to estimate the exact point that you started and finished and even when you actually followed the correct course, it can still be off in terms of it's start and end times since it is limited by the actual GPS data points that were recorded.

[Mattlikesbikes]

You can also report stuff like this and either a smarter robot or a real person at strava will look at it.