Using Calories to estimate avg watts?

[Newbie321]

As we all know strava's power estimates are not the best, for instance my avg 'watts' on the MTB is higher than on my road bike despite normally actually taking it easier on the MTB (probably due to riding a bit on tar to get to the trails)

Out of interest I decided to see if I could work out what my avg watts would be using the Calorie count generated by my Garmin with a HRM. The formula ended up being (Cal * 4.186 * .22 * 1000) / (Time in Minutes * 60), I just took the standard kj to calories formula and moved the variables around. The .22 is assuming and average metabolic efficiency, this could go higher or lower the fitter you are.

While the avg watts measure is still subject to lots of errors, it does do a better job of normalizing the effort between different bike types IMO:

 

Just thought I'd share and see what people think

[RynoVW]

Interesting method. Good way to ballpark in retrospect

[M L]

23 hours ago, Newbie321 said:

As we all know strava's power estimates are not the best, for instance my avg 'watts' on the MTB is higher than on my road bike despite normally actually taking it easier on the MTB (probably due to riding a bit on tar to get to the trails)

Out of interest I decided to see if I could work out what my avg watts would be using the Calorie count generated by my Garmin with a HRM. The formula ended up being (Cal * 4.186 * .22 * 1000) / (Time in Minutes * 60), I just took the standard kj to calories formula and moved the variables around. The .22 is assuming and average metabolic efficiency, this could go higher or lower the fitter you are.

While the avg watts measure is still subject to lots of errors, it does do a better job of normalizing the effort between different bike types IMO:

 

Just thought I'd share and see what people think

Would be interesting to compare something like this with a power meter or watt bike.

[Spinnekop]

So I am no expert and the math boffins will be able to throw in a bunch of stuff.

But Power (watts) basically has elements of torque, cadence, speed and slope in it.  
Thus your MTB stava figures will be influenced by how heavy you said your MTB is compared to how heavy your road bike is set at.
Example, to ride a 1km road with gradient of 1% at cadence 80 and speed of 30 will require a certain watts.
With all the above equal, your MTB will give you HIGHER watts (if the weight setting of the MTB on strava is higher)

That is my gut feel.

Using calories IMO is a bit of a hit and miss.  Calories will be based upon HR, correct?  This in turn is influenced by diet, tired, overworked, sleep, alcohol etc etc etc.  
 

[Frosty]

42 minutes ago, Spinnekop said:

So I am no expert and the math boffins will be able to throw in a bunch of stuff.

But Power (watts) basically has elements of torque, cadence, speed and slope in it.  
Thus your MTB stava figures will be influenced by how heavy you said your MTB is compared to how heavy your road bike is set at.
Example, to ride a 1km road with gradient of 1% at cadence 80 and speed of 30 will require a certain watts.
With all the above equal, your MTB will give you HIGHER watts (if the weight setting of the MTB on strava is higher)

That is my gut feel.

Using calories IMO is a bit of a hit and miss.  Calories will be based upon HR, correct?  This in turn is influenced by diet, tired, overworked, sleep, alcohol etc etc etc.  
 

On your Strava comment, my guess is that weight is understated. Total weight is bike + rider + all the **** they carry (for some it’s enough for another discussion).

 

On the last comment, kJ is a unit of energy. Energy is the amount of work that can be performed by force. Power is the rate that energy is moved. Calories are converted to kJ, and there is the efficiency (of lack thereof) of the body as you mentioned. So it’s always a percent of a percent, or a guess of an unknown. 
 

moral of the story… trap net.

 

[Spinnekop]

2 minutes ago, Frosty said:

On the last comment, kJ is a unit of energy. Energy is the amount of work that can be performed by force.

 

Then on a bike without a powermeter, the "energy" is derived from HR?

[Newbie321]

1 hour ago, Spinnekop said:

So I am no expert and the math boffins will be able to throw in a bunch of stuff.

But Power (watts) basically has elements of torque, cadence, speed and slope in it.  
Thus your MTB stava figures will be influenced by how heavy you said your MTB is compared to how heavy your road bike is set at.
Example, to ride a 1km road with gradient of 1% at cadence 80 and speed of 30 will require a certain watts.
With all the above equal, your MTB will give you HIGHER watts (if the weight setting of the MTB on strava is higher)

That is my gut feel.

Using calories IMO is a bit of a hit and miss.  Calories will be based upon HR, correct?  This in turn is influenced by diet, tired, overworked, sleep, alcohol etc etc etc.  
 

The math boffins are free to check my math, I only had SG not HG

The basic premise is that using the 'kilojoules = watts X seconds / 1000' combined with '4.186 kJ = 1 kcal = 1 Calorie' and 'kcal = kJ / 4.186 / .22' we work out the average watts based on the estimated calories expended during the ride. Here is an article with all the formulas and explaining them: https://stagescycling.com/en_us/content/what-it-means-watts-to-kjs-to-kcals
 

Now how accurate the estimated calories burned are is another question but at least it should be consistently right/wrong between different rides/bikes as long as fatigue etc as you mentioned are more or less in the same ballpark. So there is less dependence on bike, friction, cda etc and a consistent measure of your own HR data.

I actually exported all my garmin rides to csv and ran it on 80 rides and it was very consistent. The general trend was that on the MTB the avg watts is less, which makes sense to me as well because a lot of the time on the MTB you don't pedal when going downhill while on the road bike you do pedal. This is over the entire ride not just sections.

This would not work for normalized power for instance thus only avg.

Edited February 11, 2022 by Newbie321

[Frosty]

59 minutes ago, Spinnekop said:

Then on a bike without a powermeter, the "energy" is derived from HR?

Yes, but it's an estimation.

https://assets.firstbeat.com/firstbeat/uploads/2015/10/white_paper_energy_expenditure_estimation.pdf

[Fat Boab]

On 2/9/2022 at 7:47 PM, Newbie321 said:

As we all know strava's power estimates are not the best, for instance my avg 'watts' on the MTB is higher than on my road bike despite normally actually taking it easier on the MTB (probably due to riding a bit on tar to get to the trails)

Out of interest I decided to see if I could work out what my avg watts would be using the Calorie count generated by my Garmin with a HRM. The formula ended up being (Cal * 4.186 * .22 * 1000) / (Time in Minutes * 60), I just took the standard kj to calories formula and moved the variables around. The .22 is assuming and average metabolic efficiency, this could go higher or lower the fitter you are.

While the avg watts measure is still subject to lots of errors, it does do a better job of normalizing the effort between different bike types IMO:

 

Just thought I'd share and see what people think

Do you have any idea how Strava estimates power? (Honest question as I don't know the answer).

[Frosty]

Strava estimates it based on some known factors:

Speed
Weight
Elevation change

https://support.strava.com/hc/en-us/articles/216919427-How-to-Get-Power-for-Your-Rides

[Frosty]

P = m * g * sin(arctan(grade)) * v,
where P refers to the power required, the m is the mass of the rider and bicycle, the g is the gravitational constant which is 9.8 and grade is the slope of the hill and v is the speed for that specific segment.

Edited February 11, 2022 by Frosty

[Yyyy]

Interesting experiment but based on my experience HR data is impacted by so much more external factors that it can in some cases be a completely useless proxy for power or work done or calories burned.

Dont get me wrong, HR is a good metric for how much strain your entire system is taking but I wouldnt depend on it for a power reading.

OP, i dont know how much you investing in this sport but if its anything like most fun riders out there I would propose buying a powermeter for your bikes. Lots of fun riders with expensive bikes but claim powermeters are too expensive of an upgrade.........just saying

 

[Jehosefat]

30 minutes ago, Frosty said:

P = m * g * sin(arctan(grade)) * v,
where P refers to the power required, the m is the mass of the rider and bicycle, the g is the gravitational constant which is 9.8 and grade is the slope of the hill and v is the speed for that specific segment.

As far as I know it also makes some assumptions around other losses (rolling resistance, aerodynamics etc.) based on the type of ride/bike you select. E.g. MTB has higher rolling resistance and more wind resistance so going the same speed on a given segment will give you a higher estimated power on a MTB as opposed to a road bike.

That said, the Strava estimated power for MTB is completely off. It has no idea about the terrain (e.g. rocky, thick sand, manicured trails) all of which massively impacts the actual power number.

[Newbie321]

9 minutes ago, YaseenEnos said:

Interesting experiment but based on my experience HR data is impacted by so much more external factors that it can in some cases be a completely useless proxy for power or work done or calories burned.

Dont get me wrong, HR is a good metric for how much strain your entire system is taking but I wouldnt depend on it for a power reading.

OP, i dont know how much you investing in this sport but if its anything like most fun riders out there I would propose buying a powermeter for your bikes. Lots of fun riders with expensive bikes but claim powermeters are too expensive of an upgrade.........just saying

 

I don't care about power data, just did it to see more or less what the end result would be compared to strava. To much time on my hands between cycling

[Frosty]

3 hours ago, Jehosefat said:

As far as I know it also makes some assumptions around other losses (rolling resistance, aerodynamics etc.) based on the type of ride/bike you select. E.g. MTB has higher rolling resistance and more wind resistance so going the same speed on a given segment will give you a higher estimated power on a MTB as opposed to a road bike.

That said, the Strava estimated power for MTB is completely off. It has no idea about the terrain (e.g. rocky, thick sand, manicured trails) all of which massively impacts the actual power number.

There are other equations too. I should've posted the link to the website..

https://optimistminds.com/how-does-strava-calculate-power/

 

[mecheng89]

Remember, watts is Joules/second.