Obviously, it doesn’t account for bike/rider weight or wind speed/direction… But it DOES account for elevation, speed, and distance…
People who have seen both, is it in the ballpark? Is it consistent at all?
(I am sure it isn’t a substitute for a PM… I am really just curious.)
Gracias 
Obviously, it doesn’t account for bike/rider weight or wind speed/direction… But it DOES account for elevation, speed, and distance…
People who have seen both, is it in the ballpark? Is it consistent at all?
(I am sure it isn’t a substitute for a PM… I am really just curious.)
Gracias
I have seen it off by quite a bit, both high and low. It is rarely on the mark.
(I notice it because sometimes I accidentally have rides upload off two devices, one is recording power, the other isn’t, causing me to have to delete one)
IF it happens to be close, that’s just random dumb luck.
It’s about as accurate as your lucky numbers in a fortune cookie.
I believe it does take into account ride weight.
I find it +/- 25%, so not a total guess, but would not really base much off of that number.
Comparing rides I have done with and without my power meter, it is actually a decent rough estimate IF there is no wind, the road is flat, you are not drafting and you are tooling along at a relatively easy and steady effort level. It is wildly off if there is wind and/or you are riding in a group, like by a couple hundred watts.
If you’re climbing a steep, massive mountain and have your weight entered in Strava accurately, it should be pretty close. The physics calculations for that are pretty straightforward and will dominate the necessarily-highly-variable wind/rolling-resistance/aero calculations.
In all other situations, it’s pretty much useless.
That said, if you have a big climb nearby and want a FTP test, your speed up a steep (7+% grade) hill is approximately linearly correlated with your W/kg output. So a steep hill can substitute for a PM if you just want accurate and more-or-less repeatable fitness tests.
Not even close for me when I log a Strava ride with/without (especially something like a Crit). Might be OK for a triathlete/TT I guess since less stochastic factors.
Not useful for any type of comparison(s).
Generally not close at all. Closest for hill climbs. Horrible for group rides. Not good for flat riding in general, including TTs. Not good for downhills.
if you’re climbing a steep massive mountain when it’s not windy and you have accurate elevation measurements on your GPS device OR you happen to be in an area where Strava’s elevation model is decent (meaning the USA).
I am riding in Germany a lot these days and Strava seems to be using the SRTM elevation model which is, quite frankly, garbage. Meaning it has no idea whether I’m going uphill or downhill half the time.