Go Back  Bike Forums > Bike Forums > General Cycling Discussion
Reload this Page >

How accurate are Strava's wattage calculations?

Notices
General Cycling Discussion Have a cycling related question or comment that doesn't fit in one of the other specialty forums? Drop on in and post in here! When possible, please select the forum above that most fits your post!

How accurate are Strava's wattage calculations?

Old 11-10-16, 07:19 AM
  #1  
imperius
Member
Thread Starter
 
Join Date: Aug 2016
Posts: 36
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Likes: 0
Liked 0 Times in 0 Posts
How accurate are Strava's wattage calculations?

Hi,

I'm just wondering how accurate are Strava's wattage calculations if you have a good cycling computer and the correct data inside Strava, how do then those calculations compare to a real powermeter?

Thanks!
imperius is offline  
Old 11-10-16, 07:34 AM
  #2  
rm -rf
don't try this at home.
 
rm -rf's Avatar
 
Join Date: Jan 2006
Location: N. KY
Posts: 5,206
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 687 Post(s)
Liked 62 Times in 49 Posts
This has been covered quite often. Strava may be most accurate on long steady climbs. There, air resistance is less of a factor, so drafting and wind speed doesn't affect the calculation as much.

To search bike forums, I use "site:bikeforums.net" in the google search:

site:bikeforums.net strava power calculation accuracy

Lots of good threads on this subject!

....
This makes sense:

Originally Posted by cydewaze View Post
I think Strava is best used when used to measure improvement over previous efforts rather than to calculate actual measurements. For me, it's constantly arguing with my Garmin over things like calories and feet climbed, even though the Garmin should be more accurate on those since it knows my bike weight and heart rate.

Last edited by rm -rf; 11-10-16 at 07:41 AM.
rm -rf is offline  
Old 11-10-16, 09:13 AM
  #3  
Sy Reene
Advocatus Diaboli
 
Sy Reene's Avatar
 
Join Date: Feb 2015
Location: Wherever I am
Posts: 6,583

Bikes: Merlin Cyrene, Nashbar steel CX

Mentioned: 11 Post(s)
Tagged: 1 Thread(s)
Quoted: 3399 Post(s)
Liked 661 Times in 450 Posts
Does Strava calculate using lookup tables of power data reported by riders who use meters-- plus factoring in rider and bike weights and such?
Sy Reene is offline  
Old 11-10-16, 09:35 AM
  #4  
up on two
Senior Member
 
Join Date: May 2016
Posts: 105
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 39 Post(s)
Likes: 0
Liked 0 Times in 0 Posts
Strava is a free app and worth every cent. If a higher degree of accuracy is needed, buy whatever item you can afford to provide that information to you. In this case, a power meter.

I like Strava for keeping track of my rides, routes and mileage. If I was a more serious athlete and needed more information about my rides, then like a better bike, I would invest in better equipment too.
up on two is offline  
Old 11-10-16, 12:48 PM
  #5  
smarkinson
Senior Member
 
Join Date: May 2015
Posts: 1,003
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 332 Post(s)
Liked 13 Times in 7 Posts
Originally Posted by Sy Reene View Post
Does Strava calculate using lookup tables of power data reported by riders who use meters-- plus factoring in rider and bike weights and such?
I don't think so. Besides weight, one major factor is the accuracy of the elevation data. I have one ride where I went flat out on an uphill segment and coasted on the downhill segment immediately afterwards. As it was raining the altimeter in my GPS was giving bad data so it made the hill seem much shorter than it really was.
Because of the bad elevation data Strava thought I was doing something like 100w uphill and 300w coasting downhill.
smarkinson is offline  
Old 11-10-16, 01:30 PM
  #6  
wphamilton
Senior Member
 
wphamilton's Avatar
 
Join Date: Apr 2011
Location: Alpharetta, GA
Posts: 15,177

Bikes: Nashbar Road

Mentioned: 70 Post(s)
Tagged: 0 Thread(s)
Quoted: 2869 Post(s)
Liked 260 Times in 173 Posts
On steep hills, mGh/t is simple and accurate at low speeds. As long as Strava has the right weight it will be as accurate as its elevation estimates.

That said, if you know the the climb is, say, 100 feet and the Strava record for the ride shows it at 80 feet, I'd just multiply the power estimate (on that uphill section) by 100/80.
wphamilton is offline  
Old 11-10-16, 01:33 PM
  #7  
nycphotography
NYC
 
nycphotography's Avatar
 
Join Date: Sep 2006
Posts: 3,696
Mentioned: 18 Post(s)
Tagged: 0 Thread(s)
Quoted: 1145 Post(s)
Liked 57 Times in 40 Posts
Originally Posted by Sy Reene View Post
Does Strava calculate using lookup tables of power data reported by riders who use meters-- plus factoring in rider and bike weights and such?
What lookup table would they reference to account for the 10mph head wind I had, except for when I was following the plumber van?

How does Strava calculate accurately on my 400ft climb when today I have a headwind (which is really a cross wind except it's been channeled along the cliff wall straight into my face)?

Last edited by nycphotography; 11-10-16 at 01:36 PM.
nycphotography is offline  
Old 11-10-16, 01:34 PM
  #8  
caloso
Senior Member
 
caloso's Avatar
 
Join Date: Jul 2002
Location: Sacramento, California, USA
Posts: 39,726

Bikes: Ridley Excalibur, Gazelle Champion Mondial, On-One Pompino, Specialized Rock Hopper

Mentioned: 63 Post(s)
Tagged: 0 Thread(s)
Quoted: 2485 Post(s)
Liked 1,586 Times in 799 Posts
I have found it to be relatively accurate on a long steady climb, say Mt. Rose or Mt. Diablo. 6-7% for 10 miles. For shorter or more undulating climbs, not so much. On the flats, it's pretty worthless.
caloso is offline  
Old 11-10-16, 01:38 PM
  #9  
nycphotography
NYC
 
nycphotography's Avatar
 
Join Date: Sep 2006
Posts: 3,696
Mentioned: 18 Post(s)
Tagged: 0 Thread(s)
Quoted: 1145 Post(s)
Liked 57 Times in 40 Posts
The short answer is: If you need (or want) accurate power, or accurate calories (kj), buy a strain gauge based power meter. SRM, Quarq, Garmin Pedals, Crank arm based, PowerTap hub, etc etc etc. (But not an implied "meter" such as a PowerPod.)
nycphotography is offline  
Old 11-11-16, 01:12 PM
  #10  
Sy Reene
Advocatus Diaboli
 
Sy Reene's Avatar
 
Join Date: Feb 2015
Location: Wherever I am
Posts: 6,583

Bikes: Merlin Cyrene, Nashbar steel CX

Mentioned: 11 Post(s)
Tagged: 1 Thread(s)
Quoted: 3399 Post(s)
Liked 661 Times in 450 Posts
Originally Posted by smarkinson View Post
I don't think so. Besides weight, one major factor is the accuracy of the elevation data. I have one ride where I went flat out on an uphill segment and coasted on the downhill segment immediately afterwards. As it was raining the altimeter in my GPS was giving bad data so it made the hill seem much shorter than it really was.
Because of the bad elevation data Strava thought I was doing something like 100w uphill and 300w coasting downhill.
So this would be an instance where if Strava could use lookup tables, and using your example, it would know the power output for this segment as reported by dozens/hundreds of riders, could average this, and would apply auto-correction to your data?

Originally Posted by nycphotography View Post
What lookup table would they reference to account for the 10mph head wind I had, except for when I was following the plumber van?

How does Strava calculate accurately on my 400ft climb when today I have a headwind (which is really a cross wind except it's been channeled along the cliff wall straight into my face)?
Of course it can't accommodate these vagaries, but could Strava improve its wattage estimates by using its own database of hardware reported real-life data? OTOH if the segments you're speaking of, typically have a common wind direction if not strength (eg. wind doesn't usually come from the east one day, then the west the next, etc), then a lookup database could somewhat incorporate this trend's effect on wattage (moreso than there never being wind anywhere in any direction).
Sy Reene is offline  
Old 11-11-16, 01:19 PM
  #11  
Seattle Forrest
Senior Member
 
Seattle Forrest's Avatar
 
Join Date: Mar 2010
Location: Seattle, WA
Posts: 21,538
Mentioned: 71 Post(s)
Tagged: 0 Thread(s)
Quoted: 13381 Post(s)
Liked 4,668 Times in 2,619 Posts
It isn't the case that a stretch of road requires a specific power output, so lookup tables wouldn't work. Even if prevailing wind meant always wind and the wind speed never changed. You'd still have different riders on different bikes going different speeds with different power requirements.
Seattle Forrest is offline  
Old 11-11-16, 01:26 PM
  #12  
nycphotography
NYC
 
nycphotography's Avatar
 
Join Date: Sep 2006
Posts: 3,696
Mentioned: 18 Post(s)
Tagged: 0 Thread(s)
Quoted: 1145 Post(s)
Liked 57 Times in 40 Posts
It's like Garmin eventually improved their calorie estimates such that a 400kj ride no longer claimed 1200 calories, now it claims a somewhat less unreasonable 600 calories.

I think it was better at 1200 where anyone with any sense clearly knew it was complete BS. At 600 people might be inclined to believe it, while it is still just a BS number with only the slightest correlation with reality.

If you need or want power numbers, get a device that measures it. Period.

Otherwise you may as well be estimating your neighbors bank balance based on his house and cars.
nycphotography is offline  
Old 11-11-16, 01:48 PM
  #13  
fietsbob
Banned
 
Join Date: Jun 2010
Location: NW,Oregon Coast
Posts: 43,599

Bikes: 8

Mentioned: 197 Post(s)
Tagged: 0 Thread(s)
Quoted: 7607 Post(s)
Liked 1,302 Times in 817 Posts
Buy a power meter crankset and cross reference the 2..
fietsbob is offline  
Old 11-11-16, 02:32 PM
  #14  
Sy Reene
Advocatus Diaboli
 
Sy Reene's Avatar
 
Join Date: Feb 2015
Location: Wherever I am
Posts: 6,583

Bikes: Merlin Cyrene, Nashbar steel CX

Mentioned: 11 Post(s)
Tagged: 1 Thread(s)
Quoted: 3399 Post(s)
Liked 661 Times in 450 Posts
Originally Posted by Seattle Forrest View Post
It isn't the case that a stretch of road requires a specific power output, so lookup tables wouldn't work. Even if prevailing wind meant always wind and the wind speed never changed. You'd still have different riders on different bikes going different speeds with different power requirements.
Of course. But that's the case today. We're talking about what Strava could do to make its calculated wattage readings resemble reality more often. I'm talking about using lookup tables IN ADDITION to what Strava already does. Does Strava already use rider weight, bike weight, wheel circumference (which helps define a bike 'type') in wattage calculations as well?

The responses in this thread ie. "buy a power meter" are not helpful, and I too am interested in someone actually answering the OP's question about how much and where Strava differs from reality. 10%, 50%, etc.. off?
Sy Reene is offline  
Old 11-11-16, 02:46 PM
  #15  
Seattle Forrest
Senior Member
 
Seattle Forrest's Avatar
 
Join Date: Mar 2010
Location: Seattle, WA
Posts: 21,538
Mentioned: 71 Post(s)
Tagged: 0 Thread(s)
Quoted: 13381 Post(s)
Liked 4,668 Times in 2,619 Posts
Strava doesn't know your wheel circumference. They can't even assume 700c because they have MTB data too. Even for road rides specifically, people use a range of tire sizes and each tire has a different rolling resistance. Then you have things like aero frames and wheels.

They don't know your position (tops, hoods, drops) or whether you wore a baggy jacket that acted like a parachute or if you had a skin suit and shoe covers on.

Seattle Forrest is offline  
Old 11-11-16, 02:48 PM
  #16  
Seattle Forrest
Senior Member
 
Seattle Forrest's Avatar
 
Join Date: Mar 2010
Location: Seattle, WA
Posts: 21,538
Mentioned: 71 Post(s)
Tagged: 0 Thread(s)
Quoted: 13381 Post(s)
Liked 4,668 Times in 2,619 Posts
Originally Posted by Sy Reene View Post
The responses in this thread ie. "buy a power meter" are not helpful, and I too am interested in someone actually answering the OP's question about how much and where Strava differs from reality. 10%, 50%, etc.. off?
I agree that a better answer would be more useful. Unfortunately I don't agree that a drastically better answer is possible.

Strava's estimate is pretty close to reality when you do a long uphill, but it's further from reality when you draft in a paceline. How much % off depends on the steepness of the hill and how close you were to the rider ahead of you.
Seattle Forrest is offline  
Old 11-11-16, 03:13 PM
  #17  
Sy Reene
Advocatus Diaboli
 
Sy Reene's Avatar
 
Join Date: Feb 2015
Location: Wherever I am
Posts: 6,583

Bikes: Merlin Cyrene, Nashbar steel CX

Mentioned: 11 Post(s)
Tagged: 1 Thread(s)
Quoted: 3399 Post(s)
Liked 661 Times in 450 Posts
Originally Posted by Seattle Forrest View Post
I agree that a better answer would be more useful. Unfortunately I don't agree that a drastically better answer is possible.

Strava's estimate is pretty close to reality when you do a long uphill, but it's further from reality when you draft in a paceline. How much % off depends on the steepness of the hill and how close you were to the rider ahead of you.
Well, as a start, is there a way to get people to tell us the total watt-hours or something like this (if this is right term?) reported by a power meter over substantial mixed 40 mile+ ride, and compare to what Strava said for the same?

I thought maybe wheel circumference could make its way to Strava, as it's known by users of eg. Garmin edge gps units. Apparently this isn't uploaded to Strava though. Likewise, garmin connect account profile captures rider weight; Strava has input for bike weight. Strava could ask for bike types, wheel profile... lots of things that are just data points that could cross-reference and compare to others whose profiles of equipment, physical attributes, etc.. are the same.
Sy Reene is offline  
Old 11-11-16, 03:52 PM
  #18  
Seattle Forrest
Senior Member
 
Seattle Forrest's Avatar
 
Join Date: Mar 2010
Location: Seattle, WA
Posts: 21,538
Mentioned: 71 Post(s)
Tagged: 0 Thread(s)
Quoted: 13381 Post(s)
Liked 4,668 Times in 2,619 Posts
Originally Posted by fietsbob View Post
Buy a power meter crankset and cross reference the 2..
There isn't an easy way to do this, because Strava won't estimate your power when you've measured it.

You'd have to:
  • edit your file to remove the power data,
  • send the edited file to them,
  • download the ride file after they've processed it,
  • and then compare them on a second-by-second basis.
And even then you couldn't generalize to all of Strava; what you'd get would be specific to the ride you did.
Seattle Forrest is offline  
Old 11-11-16, 03:55 PM
  #19  
Seattle Forrest
Senior Member
 
Seattle Forrest's Avatar
 
Join Date: Mar 2010
Location: Seattle, WA
Posts: 21,538
Mentioned: 71 Post(s)
Tagged: 0 Thread(s)
Quoted: 13381 Post(s)
Liked 4,668 Times in 2,619 Posts
Originally Posted by Sy Reene View Post
Well, as a start, is there a way to get people to tell us the total watt-hours or something like this (if this is right term?) reported by a power meter over substantial mixed 40 mile+ ride, and compare to what Strava said for the same?
If you'd like, I can send you some FIT or GPX files with power measurement in them and my weight that day. You can analyze them however you like, and maybe come up with something clever. Golden Cheetah can open FIT files and let you copy/paste the data into Excel, and I have a SQL query for reading GPX and TCX data.
Seattle Forrest is offline  
Old 11-11-16, 04:10 PM
  #20  
nycphotography
NYC
 
nycphotography's Avatar
 
Join Date: Sep 2006
Posts: 3,696
Mentioned: 18 Post(s)
Tagged: 0 Thread(s)
Quoted: 1145 Post(s)
Liked 57 Times in 40 Posts
Look, this is your big idea, so why not have a look at some actual data and see if you can really make any sense of it.

Pick any segment of a reasonable length that is neither dead flat, nor a sustained climb. Here's one (Closter dock to GWB): https://www.strava.com/segments/11059501?filter=overall

This is a basically flat to descending segment. Pick any two riders with very close times, one with a power meter (example: Andrew McGee, 14:34 @ 274w) one without (ex: Michael Sambrano, 14:36 @384 est watts). While you're there, look at some of the other results.

I will say this... it appears their estimates are way high pretty much across the board!

(I'd bet cash that all of the these top results are TEAM/GROUP rides w/ lots of drafting, and this power difference is EXACTLY the kind of error that simply can't be estimated around.)

This pretty much illustrates why they can't improve their estimates.
nycphotography is offline  
Old 11-11-16, 04:16 PM
  #21  
nycphotography
NYC
 
nycphotography's Avatar
 
Join Date: Sep 2006
Posts: 3,696
Mentioned: 18 Post(s)
Tagged: 0 Thread(s)
Quoted: 1145 Post(s)
Liked 57 Times in 40 Posts
And 4 of the top 5 (along with several more farther down) are all on the same day... probably a team coming back down 9w and almost certainly with a sweet tailwind.

Notice the estimates in that group all around 400w to everyone with a meter around 275w (except for lazy Ira sucking on the back of the line at 239w lol)
nycphotography is offline  
Old 11-11-16, 04:40 PM
  #22  
wphamilton
Senior Member
 
wphamilton's Avatar
 
Join Date: Apr 2011
Location: Alpharetta, GA
Posts: 15,177

Bikes: Nashbar Road

Mentioned: 70 Post(s)
Tagged: 0 Thread(s)
Quoted: 2869 Post(s)
Liked 260 Times in 173 Posts
Originally Posted by Seattle Forrest View Post
There isn't an easy way to do this, because Strava won't estimate your power when you've measured it.

You'd have to:
  • edit your file to remove the power data,
  • send the edited file to them,
  • download the ride file after they've processed it,
  • and then compare them on a second-by-second basis.
And even then you couldn't generalize to all of Strava; what you'd get would be specific to the ride you did.
Or, run the Strava app on your smartphone and compare that to the power meter.
wphamilton is offline  
Old 11-11-16, 05:08 PM
  #23  
Seattle Forrest
Senior Member
 
Seattle Forrest's Avatar
 
Join Date: Mar 2010
Location: Seattle, WA
Posts: 21,538
Mentioned: 71 Post(s)
Tagged: 0 Thread(s)
Quoted: 13381 Post(s)
Liked 4,668 Times in 2,619 Posts
You could do that too. But you might not learn as much because of the way the numbers jump around.
Seattle Forrest is offline  
Related Topics
Thread
Thread Starter
Forum
Replies
Last Post
RShantz
Road Cycling
62
05-13-19 03:26 PM
FlamsteadHill
General Cycling Discussion
81
08-25-18 09:08 AM
Baldy1953
General Cycling Discussion
12
08-02-18 04:45 PM
Caymandiver1
Training & Nutrition
15
06-05-16 11:45 AM
chaadster
Road Cycling
30
08-02-15 12:29 PM

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off


Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service - Do Not Sell My Personal Information -

Copyright 2021 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.