In this post I’m going to look at how Strava calculates the power generated over the course of a ride, specifically where the rider (like me) doesn’t have a device attached to the bike for measuring power.
Dedicated power meters are still an expensive addition to your bike. Is it worth spending that extra money when Strava can provide the data for free?
Why Do We Care About Power?
The Power Of Love, A Force From Above – Frankie (Goes To Hollywood)
The use of power data as a training and racing tool seems to have grown in profile in recent years.
In the pro ranks (and in some amateur ones), time trialists use power meters to help them spread their effort over the full course of a race without blowing up. Team Sky use power meters in order to ride tempo up the steep passes of the Tour De France, preventing attacks from other teams (and, according to critics, sapping interest from the race).
On the consumer side, power meters are starting to come down to a price that’s within reach of amateur riders (admittedly, amateur riders with plenty of disposable income). Now that the likes of Garmin and Polar (but mainly Garmin) have easy-to-install pedal-based meters available (as opposed to cranks and rear hubs), you can only see them growing in popularity.
But do you need to spend £’000s on a power meter when Strava provides the ‘data’ for free?
To answer that question, we need to look at how Strava calculates its power data, and how accurate it is for you as a rider.
Wherein The Grimpeur Attempts To Explain How Strava Calculates Pow’r
Strava calculates the overall power output of the rider and bike ‘system’ rather than the wattage generated at the cranks, pedals or hub (as would be the case when you use a power meter).
It takes your measured performance (speed, distance, time), adds in some environmental factors (some actual, some assumed) and comes up with wattage numbers over the course of your ride.
Since the wattage is that of the whole rider/bike system, Strava’s data is most accurate when it knows more about the rider and the bike (e.g. rider weight, bike weight, type of bike).
Which ‘Components’ Does Strava Use To Calculate Power?
I am not a physicist (I’m a historian), so I don’t know whether ‘component’ is the correct word to use. It just felt a bit better than ‘things’.
Anyhoo, Strava uses the following, er, things in its power calculation formula:
- Wind resistance
- Rolling resistance
- Gravity
- Acceleration
What Strava Doesn’t Know (And What You Should Tell It)
Each of the components mentioned above has it’s own formula (i.e. to calculate the power required to overcome that particular thing).
Within the wind resistance calculation, for example, Strava doesn’t know the environmental conditions under which you were riding (or at least they don’t know at the level of detail required). So, according to the website, Strava assumes no environmental wind conditions and a constant air temperature of 15 degrees (as far as this relates to air density).
In addition, there may be things you haven’t told it.
The drag coefficient within the wind resistance calculation is based, in part, on the type of bike that you’re riding (which you set when you ‘Add Bike’ under Settings > My Gear). Presumably it makes an assumption on the drag caused by an average rider in an average position on the bike.
Similarly, if you’ve added your weight and that of your bike, these are used to calculate the power required to accelerate the bike and the power required to overcome gravity as you soar gracefully up that mountain road.
How Does Strava Power Data Differ From Power Meter Data?
As discussed above, Strava takes the measured output of you and your bike together and then backs out an estimate of the power you employed. Power meters measure the actual power produced at a specific point on the bike (i.e. where the measuring device is located).
Power meters do not need to take into account terrain data, weather conditions and how aerodynamic you are, all of which Strava has to contend with.
If it’s really windy, the elevation data being used is dodgy and you’ve decided to ride with drag-racing style parachute attached to your seat post, then Strava’s power data is unlikely to be very accurate.
That said, Strava claim that ‘in most cases our watts number are very close to the numbers provided by a Powertap or SRM’.
So there you go then, question answered….
So Do You Need a Power Meter?
Speaking for myself, a power meter is not top of my wishlist (well, it might be, but what I mean is that I ain’t going to be buying one any time soon).
Whilst I’m sure I’d find capturing the data interesting, to be spending that sort of money, I’d have to be using it to train properly. Many of the benefits of training with power can be achieved by training with heart rate zones. If I can prove to myself that I can use heart rate data consistently, maybe that will be the time to purchase a power meter (and perhaps by then the prices will have fallen).
If you do want to train with power then you need to be able to view your wattage output in realtime whilst riding. As far as I’m aware, Strava doesn’t do this, even if you have your iPhone mounted on your handlebars (correct me in the comments if I’m wrong on this). For this you’d need to buy a power meter.
Wes Fallan liked this on Facebook.
Good article! I do similar calculations, and make similar assumptions for the trains we build, to check we have the righht amount of installed power to meet the time table. We would call it “power at wheel”. And the “things” are know as “parameters”, but “things” works as well!
Rob – I thought a train modelling enthusiast had stumbled upon my site… then I realised it was you!
Glad that ‘things’ works for you. Maybe a career in engineering beckons…
This is exactly why myself and two colleagues (software engineers all three) spent a bit of time recently working on this:
https://github.com/mjervis/strava-centurion
It takes the TCX file from your garmin ride, and performs physics based calculations on the ride data given the input of your weight, bike weight, drag area and coefficient of rolling resistance.
There are some sensible defaults (ok, my metrics, based on a Road Bike with 23mm slicks riding on average UK tarmac) in the app, but we found some nice research somewhere with different drag coefficients that we use as a look up table when we switch bikes, or take a more “touring” riding position.
There are some issues with the current build around the output from a 705 (each garmin varies a little in the content) as the guy who last checked in made sure it worked with his 800 and broke it for my 705 🙁 works for a 500 currently I believe (conveniently we each have different garmins).
We did it as with a Strava Premium subscription and a real wire tap, you get loads of interesting graphs and figures (http://app.strava.com/activities/53834726).
But, as your weight varies over the ride (water bottles empty, gels swallowed, toilets visited, cakes eaten) and the wind changes direction etc it’s all an estimate.
We also get some extreme spikes in the power as glitches in the altitude make it look like you rode up a 90% gradient for a few seconds etc.
But interesting to look at, if you view it in context.
(oh and you can set the days temperature too, which will adjust the air density and get you “more accurate” inaccurate estimated figures :))
Thanks Mike. I wonder if Strava are doing the same work in house? If not, maybe you should see if they’ll buy it off you (or maybe one of Strava’s competitors would).
How do I use it from the github site (or are we not meant to)?
@Andrew, You can view the source code on Github, well I guess you won’t need it considering it was almost over half a decade ago
And you can play physics here:
http://bikecalculator.com/
Wes Fallan liked this on Facebook.
Sorry, but just more “penis measuring devices”! Get out and ride, enjoy yourself, and leave the distractions and essentially meaning data behind. Unless you are planning on becoming a professional none of this is worth your time and/or money. I do most cycling with all this very “specialized” equipment: a bike; a body; cycling helmet, shorts and shoes. Gloves if cold. Everyone stopped being sucked in my by marketing that makes you think this is necessary, the cycling world would be a better place. No talk of ridiculous “data” and strava BS. Please, it’s time to get real and do some cycling instead of over “anal”izing every bloody thing. I can’t even have a pleasant conversation with someone without them bringing up the trivial aspects of cycling that don’t apply to recreational/amateur cycling. Thankfully my bike shop/collective are on the same page and regard bike/equipment maintenance and safety as the important aspects of cycling; yesterday, today, and tomorrow. Did I read that the blogger isn’t much on maintenance; it’s more important to your cycling performance, safety, and enjoyment of cycling than any of the superfluous topics generally spoken about in cycling blogs, forums, etc. Do you all walk around with a power meter, heart rate monitor, aerodynamic clothing and what not. Do want “cyborg” implants as well! If can’t enjoy life without it I feel sorry for the lot of you all. I believe I read somewhere on the blog site that this is primarily for “beginners”. Not necessary! What is necessary is how to stay alive in today’s cycling environment with many dying on the streets; thank you Strava, another one bites the dust! Cycling for over 50 years, able to decimate those younger than myself in many instances, as well as myself being left behind by older riders; none of us having all this crap strapped to themselves and/or bike. I have not competed professionally, but I have competed in various events. Always worn a cotton (very comfortable) t-shirt and riding a steel bike at least 10- 15 years “behind the times” relating to the technology when I first started “serious” riding; riding that same bike and component group for the last 35 years (maintenance!, not meters!). Now I have to deal with this “mamil” stupidity because of all the foolish need to make either an impression to oneself and/or others. Makes me want to vomit! Sorry for the reality check, but it’s long overdue. My “power meter” and “heart rate monitor” are in my brain; are yours not functioning? I prefer to be a rider, not a robot and have a much higher level of self accomplishment and satisfaction after anything I do because there is no need to waste time analyzing that which does not provide a truly better life. Go ahead everyone, take your shots at me; but in your heart you know the truth (or maybe your “reality meter” needs a new battery).
Respectfully,
Unencumbered human
Okkkaaaayyyy. You do *sound* a bit encumbered though.
I have to agree for the most part.
It all made a modicum of sense until it fell into the old man uncanny valley.
You lost me at cotton t-shirt. At no time in history have athletes worn sweaty, irritating cotton t-shirts. What a rookie move. Cyclists have always worn wool or synthetics. Going back 50 years I doubt you’ll find any cotton jerseys. Gross. Uncomfortable.
While I’ll agree that metric obsession is worthless unless you’re competing, it helps to know your distances and exertion, despite also being kind of fun.
But clearly fun left the picture some time ago. I’m not a young man, but I’m here to tell you: it’s not too in late to have a good time. It’s not too late until, of course, it is too late. Better get on that.
Pardon, it should have read “meaningless data”!
Nowadays, if you have a recent Garmin computer on your bike you can install an app (from the Garmin IQ store) that will give you an estimate of power output, supposedly using the same methodology as Strava does but in real time.