Accuracy of Strava power estimates.
 

How accurate are the power estimates listed by Strava?
Does anyone here have experience with the estimates vs. actual power-meter data?

Thanks

Yes. They're very poor.

Coast down a hill and tell me what Strava reports.

And then ride into a head wind.

I can't tell you how accurate they are versus actual power meter but I do know your altimeter needs to be working spot on to stand any chance. And rain really stuffs up your altimeter.

I finished ninth on this uphill segment (I'm crap up hills) when I was really trying but strava says I was only using 175 watts. I presume this was because my altimeter was acting up due to the rain so under reported my elevation gain (22m for this ride versus the usual 49m of the climb). Interesting that others with power meters with similar times were doing about 340w.

https://www.strava.com/segments/3439519?filter=overall

On the flipside coming down the other side of the hill (which I didn't pedal at all) Strava recorded 322w (and a 5m elev drop).

Fairly accurate for steep hills. Quite poor for everything else.

I cant really say anything as to the accuracy of the power on strava as i dont have anything to compare it to. Other then I'm sure I'm not putting of 1600 watts on sprints sometimes like it says. I did notice a huge discrepancy with it elevation gain. On my profile it shows my biggest climb as a being 286 ft. I know i haven't climbed a hill that size because i know where it picked that up at by clicking on the number on my profile. I review that segment and its only a 12 ft gain in elevation on that entire road.

I think to average for a ride might be ok, but you can't trust small parts much as it relies on often faulty elevation profile data. You also have to have your correct weight entered, and yeah it wouldn't know about wind.

many people dont have rides taht consist solely of downhill coasting

They aren't accurate.

There are 8 replies and 3 of them mention bad elevation data in Strava causing the power estimates to be way off. Are people not running the ele corrections, or are they really bad? (I'm not on Strava so I have no idea.)

Who cares? If it can't get that part right, why would you trust anything else?

Well, on some group rides that I've been on elevation for the riders who upload to strava is all over the place. Some will get, say, 2000 feet of elevation gain and some will get half that with some getting whatever in between. And it doesn't seem to matter if you are using a garmin or not, although the phone app people's elevation is usually on the lower end of the scale.

One hill which I'm on frequently, Strava thinks that my rides are 240-290 watts going down hill, and 180-200 watts going up. But I'm taking it easy on the descent, coasting parts of it, and killing myself on the return uphill. My weight is entered correctly, myself and my bike. I could fudge the weight to make it come out more reasonably, but whether that's really where Strava is miscalculating who knows? Regardless, the power estimate is way off.

Strava power data is for entertainment purposes only.

because downhill is a fraction of peoples rides and an equally insignificant fraction of their attentiveness to power measurements.

over a typical ride distance and profile the measurements converge.

all the 50-150km group rides ive been see consistent time, speed, and profile reading, whether reported from garmin, android, or iphone. and being that those (w system weight) being the key factor in power reading (w assumed average CD/WR & RR) , its decently good when compared to power meters*...

...which also arent infallible....

This is older but still relevant and interesting:

How Accurate Is Strava as a Power Meter? A Retest on a Long Ride with Wind and Hills | FreeTriSpeed.com

If you plug in your strava information (rider weight, bicycle weight and so forth) to this calculator and then plug in a specific uphill segment info you will get a number very close to your strava power estimate for that segment.

Bike Calculator

It's not that the strava power estimate is accurate or inaccurate, it's that people use it for the wrong thing. It's a repeatable number than you can compare to other efforts, assuming the other variables remain the same. But in the end it's just a formula based on weight, grade and speed.

LOL. a calibrated power meter is accurate to < 1-2% or so, which is plenty accurate for what cyclists are doing.

A strava guesstimate is accurate to within 100% or so. Doesn't know wind, doesn't know downhill, doesn't know if you're pedaling, doesn't know if you're in a paceline... about the only time it's passably accurate is if you're going up an extended, reasonably steep hill. Other than that, it's reliably wrong.

because your typical ride isnt a 10 minute tucked-train downhill bomb with a 40mph wind , but a loop that exposes the rider to a variety of slopes and wind vectors, i introduce you to the wonderful concept of averaging

its "guesstimate" is closer than anything you can muster from your head. i.e., it knows atleast more than you do.

False, because I actually have a power meter. No need to guess.

So you average a bunch of bad data and somehow that's supposed to be relevant for something? Do you use it for intervals or pacing yourself on a TT, or do you just sit around at the pub when you're done bragging about your mighty strava power numbers?

Seeing as how I don't have a power meter, and therefore cannot get feedback on current power output, I have no desire to find out historic power data as I review my ride in Strava, since it is not relevant to my training.

i removed computers from all my bikes because i dont obsess over the stats. but it is occasionally amusing to peek at post-ride. and pleasantly amusing when it matches up with pm-equipped ride mates.

your wholesale dismissal of its accuracy reveals ignorance of the empirical evidence and/or the underlying physics. speaks more about you than it does strava.

enjoy analyzing your downhill segments

So inaccurate that it shouldn't even be a "feature."

You clearly don't understand how it all works.

"amusing when it matches up with pm-equipped ride mates" <- This quote says it all.

I didn't "wholesale" dismiss it either. Your reading comprehension... lacks.

I agree that Strava power numbers are occasionally amusing. :thumb: