You are currently viewing the message boards in:

# Cycling efficiency for different riders (it varies a lot)

A lot of people trust the absolute calorie burn estimates provided by bicycles (stationary or moving) that have a power meter of some sort. But, the total calorie burn of the rider depends on knowing the rider's efficiency, which is, on average, about 20%. This means that you (or your bike computer) must take the energy expended on actual pedaling and multiply it by 5 to get the number of calories consumed by your body.

Many people think this efficiency is pretty much the same for everyone. One reason they think that is the title of the paper "No Differences in Cycling Efficiency Between World-Class and Recreational Cyclists," Mosley, et. al, Int. J. Sports Med (2004). (And other works based on it.)

I've posted this before, but the problem with this paper's title is that it really means that the average efficiency of each cohort ("world-class" or "recreational") is about the same. But the paper also shows that there is wide variation between individuals making up the entire study.

For illustration, here's a figure from the paper that shows that both the "GE165" (the efficiency factor) and the VO2max vary quite widely from person-to-person (each individual being a single data point). The point of the figure is to show that the two metrics are pretty much uncorrelated, but it also shows that the GE165 varies from 14 to 22% for people tested.

Note that the "GE165" is defined in the paper as

((Work Rate [W]/Energy Expended [J/s]) * 100%, as measured for an individual at a power output of 165W[/quote]

Here's a link to the paper

Comments welcome. (Particularly if I'm misinterpreting something! I am not an expert in this field.)

## Replies

• Member Posts: 10,457 Member
Human efficiency turning chemical energy from fats and carbs into watts ranges from about 18 to 21 %, or something like that. The window is less than 5 percentage points wide. According to every lab that's every studied it. That's why people have so much confidence in the calorie conversion.
• Member, Premium Posts: 18,551 Member
You are confusing / misinterpreting two separate things OP.

Converting power output to calories is extremely well studied, understood and has very little variation between most people.
Even cycling coaches have moved away from trying to teach people the perfect/mythical
souplesse as they have realised actual performance and power production between stamp/stamp style and a long, smooth power pushes is negligible.
Even at my highly technical bike analysis and bike fit the guy commented on my pedalling style and said "nice and smooth, shame that it doesn't help you produce more power, try just concentrating on the push". Harumph!!!

Your study is talking about a variation between people's maximal oxygen uptake capacity and their power production. It is NOT comparing power and calories.
Quite apart from few people test and know their VO2 max it also doesn't matter for using power measurements to make very accurate net calorie estimates. An individual with an extraordinary high VO2 max and someone with an ordinary CV system producing the same power (your study's 165w) are burning virtually the same net calories.

The range of cycling efficiency is more like a range of 20 - 25% which is why simply using 24% allows the complex formula to be simplified as a KJ is 0.24 of a calorie).
It also means that assuming a high efficiency rating will mean the estimate for most will be most likely an underestimate.
edited May 12
• Member Posts: 18,903 Member
This is different efficiency than the often stated 20-24% for muscle calorie burn turned into mechanical energy of torque.

muscle efficiency is what is normally being referenced with that narrow range, and that applies to any work a muscle is doing - cycling just happens to make it easier to measure the output power to backtrack to the input power (or output from the muscle).

Formula's exist for running too, using the same concept. Much harder to measure the force to backtrack on though.

But that article is discussing cycling efficiency in general, or specifically. Not muscle efficiency that is often referenced by those touting power meters.

I'd be curious to take a sample person, work back to their VO2 used (not per kg), and get calorie burn from it for that 165 watt effort.
• Member Posts: 10,457 Member
There is such a thing as running economy, a lot of it is because runners move freely in 3 dimensions. I can waste more calories bouncing up and down too high without getting enough forward motion from that energy I spend fighting gravity. Stuff like that.

On a bike, you're seated, you're turning your legs in a 2D circle (no right or left) that's always the same size. Mine's always a 175 mm radius because that's how far my pedals are from the bottom bracket. We just don't have as much opportunity to waste energy on a continual basis. Example: I did a ride last weekend, 2 hours 41 minutes, Garmin says 3 minutes of that was standing. (Strain gauges in the pedals.) That's typical, road cyclists stand occasionally but it's a negligible slice of the pie. Things are probably different in MTB, but it's kind of a moot point because those people don't use power meters. We have a different kind of economy where you can use energy that doesn't make it to forward motion, but it's a different thing because energy that doesn't make you faster still uses calories.

I'm just some guy on the internet, what's important is that this has been studied extensively with a lot of high quality research and the findings keep replicating every time they test it.
• Member Posts: 10,457 Member
For this conversation, it's important to be aware of net vs gross calories.

If I just sit on the couch, the power meter's going to say zero.

The basically 1:1 ratio of kJ to kCal still holds true, but be aware of apples and yummy oranges. I lose weight faster than predicted in the summer when I bike a lot because my BMR just disappears from the numbers since MFP assumes it's included in exercise that gets synced in. It adds up if you're riding an hour a day.
• Member, Premium Posts: 2,815 Member
sijomial wrote: »
An individual with an extraordinary high VO2 max and someone with an ordinary CV system producing the same power (your study's 165w) are burning virtually the same net calories.

I don't know. The GE165 listed in the study stands for "Gross Efficiency," meaning, I think, it's the number we're talking about. For this study, it's measured at 165W output for every individual. Ignoring the VO2max measurement and looking just at the spread of GE165 in that scatter plot, you see a significant spread between the individuals in the study. I can't see any other interpretation (although the paper does not mention it).

Here's another study of Gross Efficiency in male and female elite cyclists:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3761728/#:~:text=It has been suggested that,cost and level of energy

It shows a pretty big difference between the males and the females (mean male GE = 19.9%, mean female GE = 22.5%). Also in that study, the standard deviation inGE for elite male cyclists is 1.8% @ 150W (this is from only 13 males of about the same age. Mean GE of 19.9%). That's a pretty big variation for such a small homogenous group.
• Member, Premium Posts: 18,551 Member
sijomial wrote: »
An individual with an extraordinary high VO2 max and someone with an ordinary CV system producing the same power (your study's 165w) are burning virtually the same net calories.

I don't know. The GE165 listed in the study stands for "Gross Efficiency," meaning, I think, it's the number we're talking about. For this study, it's measured at 165W output for every individual. Ignoring the VO2max measurement and looking just at the spread of GE165 in that scatter plot, you see a significant spread between the individuals in the study. I can't see any other interpretation (although the paper does not mention it).

Here's another study of Gross Efficiency in male and female elite cyclists:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3761728/#:~:text=It has been suggested that,cost and level of energy

It shows a pretty big difference between the males and the females (mean male GE = 19.9%, mean female GE = 22.5%). Also in that study, the standard deviation inGE for elite male cyclists is 1.8% @ 150W (this is from only 13 males of about the same age. Mean GE of 19.9%). That's a pretty big variation for such a small homogenous group.

Here's a personal example of cycling efficiency:
I was naturally a slow cadence rider (60rpm) but trained myself to ride at more efficient cadence (average 85rpm) which resulted in performance improvements - I could produce more power for longer so my FTP improved.
Note that's not metabolic efficiency, I'm making more power for longer and burning more calories.

Look at Fig 4
"The metabolic cost of producing a given mechanical power output is shown in Fig.4. There were no significant differences between groups at any mechanical power. "
• Member, Premium Posts: 2,815 Member
sijomial wrote: »
Look at Fig 4
"The metabolic cost of producing a given mechanical power output is shown in Fig.4. There were no significant differences between groups at any mechanical power. "

I think this is where the paper is most misleading. It's the statement "between groups" rather than "between individuals." In other words, the means of each group was very close although the individuals varied significantly as shown in Figure 1.
• Member Posts: 10,457 Member
FYI

Here is a good explanation of the science. 🙂

• Member, Premium Posts: 20,374 Member
The groups differ. Within each group, there's individual variation. While I'm not seeing all the data I'd like, in order to support this assertion (in the first-linked study in the OP), it *looks* as if the variability within each group (SD) is pretty similar (reasonably proportional to the values being measured, which of course differ between the groups, so not necessarily numerically identical). IOW, I'm not convinced that the scatter around the trend line tells us anything meaningful, from a calorie-estimation standpoint.

Still, the standard in calorie-counting-world ought to be "close enough to be practical" . . . which using average efficiency numbers *is*, IMO, even if there's variation around the mean.

Heck, there's variation around the mean for BMR/RMR**, and since that's a bigger deal for most people (arithmetic magnitude) once one uses a "multiplier of BMR/RMR" kind of a approach to calculating base goal . . . I don't see the point of Joe or Jane Casual-cyclist stressing about the role of an efficiency assumption that affects a few percentage points of something that for most (short of y'all admirable distance/endurance riders) is some small fraction of maybe (generously) 25% of the calories in one's average day.

Me, I ride my slow li'l ol' lady distance/time on the hybrid; on a really good day it's *maybe* 20% of my TDEE . . . meh, mathematically, if the efficiency adjustment is off.

But I'm not a statistician, so I could for sure be missing something.

**https://examine.com/nutrition/does-metabolism-vary-between-two-people/
• Member Posts: 10,457 Member
@AnnPT77 is right: on a practical level, it doesn't matter very much to most people. And as a rule (with exceptions) the people who can and do do enough exercise for it to be a significant part of a weight loss journey don't need to lose weight. I think @Jthanmyfitnesspal started this thread out of intellectual curiosity.
• Member, Premium Posts: 2,815 Member
I agree with @NorthCascades and @AnnPT77 that these are all just estimates. My main point is that, even with a (pricy) power meter, your knowledge of your calorie burn is still pretty inaccurate. Of course, you're no better off when running or (even worse) swimming. And don't get me started about some other activities, such as mowing the lawn. You just take your best guess and watch your weight.
• Member Posts: 10,457 Member
Calories from a power meter can be off by +/- 2.5%.

In the scheme of weight loss, you'd have to burn more than many people eat in a day for it to be off by an Oreo. 🙂

Also for context and clarification, the reason people spend \$\$\$ on power meters and make their bikes heavier in the process isn't calories. Way too much to spend on that. A power meter is like the numbers printed on weights. Using a bike for fitness purposes, in a race, or in an endurance event, that information can be valuable. It helps you set realistic and useful training targets, see your progress, and manage the strain you're taking on. It also lets you math your way into useful aerodynamic info, if you're an amateur racer you can't afford wind tunnel time but you can use Aerolab with speed and power data collected with a clever protocol to address other variables.
• Member Posts: 18,903 Member
Something else I meant to comment on regarding that graph and some comments in the article - pointing out 2 different types of efficiency are being talked about here when using power meters.

The graph is including weight - VO2 ml/kg/min.

Calories from formula using watts/kj does not include weight.

2 riders with different weight but same VO2max for ml/kg/min are going to have different total VO2, and therefore different calorie burn from the indirect calorimetry they used.

So the fact the graph shows watts/joules (no weight) plotted against an measurement that does include weight - throws it off for me.

That graph is indeed showing cycling efficiency as it's talking about - but again that isn't the efficiency of the human motor be turned into torque/watts.

I want to dig into that paper more to see if they give enough figures to work out a partial graph without weight involved.

Something about 2 riders supplying the same volume of oxygen per kg, but very different total volumes to their potentially 2 very different amounts of muscle mass, that are both supplying 165w of power though - something isn't sitting right.
• Member, Premium Posts: 18,551 Member
I agree with @NorthCascades and @AnnPT77 that these are all just estimates. My main point is that, even with a (pricy) power meter, your knowledge of your calorie burn is still pretty inaccurate. Of course, you're no better off when running or (even worse) swimming. And don't get me started about some other activities, such as mowing the lawn. You just take your best guess and watch your weight.

Within a few percent for net calorie burns is pretty damn accurate. Assume my 200,000 annual cycling calories are in fact under-stated by 2.5% - that's a variation of just 5,000 over an entire year or less than 14 cals a day.

But you do need to understand that the graph you seem to think is plotting net calorie burns against power produced actually isn't. Which is why the x and y axis aren't labelled watts and net calories.

• Member, Greeter Posts: 2,233 Member
heybales wrote: »
Something else I meant to comment on regarding that graph and some comments in the article - pointing out 2 different types of efficiency are being talked about here when using power meters.

The graph is including weight - VO2 ml/kg/min.

Calories from formula using watts/kj does not include weight.

2 riders with different weight but same VO2max for ml/kg/min are going to have different total VO2, and therefore different calorie burn from the indirect calorimetry they used.

So the fact the graph shows watts/joules (no weight) plotted against an measurement that does include weight - throws it off for me.

That graph is indeed showing cycling efficiency as it's talking about - but again that isn't the efficiency of the human motor be turned into torque/watts.

I want to dig into that paper more to see if they give enough figures to work out a partial graph without weight involved.

Something about 2 riders supplying the same volume of oxygen per kg, but very different total volumes to their potentially 2 very different amounts of muscle mass, that are both supplying 165w of power though - something isn't sitting right.

I looked at it in the same light. The study essentially proves that to produce X amount of power requires Y calories. It doesn't really go into the total calories that the humans needed to produce X amount of power, and as such doesn't really go into the GE.

Similar to studies that often correct for fat free mass when comparing numbers between one group and another, they prove that net work has a specific cost, but the variations in humans alter the gross cost of the work.

Being that they also set an RPM range, they essentially proved that they can take a group of riders, get them to apply the same torque at the same RPM, and come up with the same power.

I don't cycle enough to justify power meters, but likely even if I did I wouldn't buy them. In the end it most of this fitness related stuff is down to best estimates, and adjust as needed. Even if a person manages to buy very accurate meters, the gross cost still isn't an exact.

Any time anything uses an average, things are not always point blank as those averages skew the actual data points.
• Member Posts: 10,457 Member
robertw486 wrote: »
I don't cycle enough to justify power meters, but likely even if I did I wouldn't buy them. In the end it most of this fitness related stuff is down to best estimates, and adjust as needed. Even if a person manages to buy very accurate meters, the gross cost still isn't an exact.

Unless they have Bill Gates money, no one buys power meters to count calories. That's not what they're for. Being able to tell you how many calories you've burned is an interesting side effect of how they work, but not worth \$300+ for most people. I'm getting this weird feeling that some MFP Mythology is being created in this thread. 🙂 Power meters are a training tool, they're useful because they measure intensity in a repeatable and precise way. The data they produce has other uses including frankly astonishing calorie accuracy. Analogy: a car has lots of uses but nobody buys one to store things on the back seat. 🙂
• Member Posts: 10,457 Member
I've been flirting with a training plan this season, usually I'm less structured about this stuff and fill in the gaps as I see them. Long story, but I'm giving it a try.

A few days ago my workout was sprint repeats. 🤮 As an interval workout. It had me warm up for 20 minutes at 200 watts, then a series of 10 seconds @ 520w followed by 20 seconds active recovery. 😳 🤬

That is why people buy power meters. Without a PM there's no way to prescribe a workout like that and there's no way to know how closely you achieved it. Terms like HIIT and really hard are too vague. Heart rate responds too slowly to use for short intervals, and for something like this it's too affected by hydration and weather. Pace works running on flat ground but is almost meaningless on a bike. But 520w x 10' has a specific meaning and is hopefully pretty close to optimal for where I was a few days ago.

If you're looking to build strength, you want to know how many kgs you're lifting, if you're looking to build fitness or maintain it with age, a power meter is like the numbers printed on the weights.
• Member, Premium Posts: 2,815 Member
I am really amazed that so many of you believe that the efficiency factor (GE) is constant between individuals "to 2.5%" when measured data for GE is staring you right in the face from my original post. Yes, it's plotted vs. VO2max, but don't let that confuse you. The range of measured GE values presented in that little study is from 14.5% to 22.5% for different individuals.

The power measured by the bicycle power meter is probably pretty good (I'd think it could be measured to within 2.5%, for example), but the conversion to calories burned by the rider is much less certain, particularly among casual riders.

And I think that is a pretty intuitive fact: if you put an untrained and heavy person on a bicycle on a flat course, there is no way they would have the same efficiency as a much lighter trained person for the very reason that their legs are heavier and take more work to pump (regardless of power applied) and they probably don't unweight the rising pedal very well. All this lowers their efficiency.
• Member Posts: 10,457 Member
Do you really think you found something that tens of hundreds of sports scientists missed while researching exactly this?? 🤯