Bike Power vs Run Power

I’ve had power meters on my bikes for a couple of years now and have just about got to grips with them (not sure if I’m any fitter/faster because of it, mind you). I recently bought a new running watch - a Coros Apex Pro - which has a wrist-based power meter. On a typical bike ride (easy-moderate pace) I average about 170-180 watts NP. But on a run of similar duration and perceived effort, I’m hitting about 250 watts.

I figure this is for one of three reasons:

  1. I’m a stronger running than I am a cyclist - no surprising given that I do more running and most of my “proper training” is running not cycling (or my actual effort is higher when running and my perception is off).
  2. The running power is based on an algorithm using HR, pace, GPS data and other things to calculate power whereas the bike power uses a physical load sensor. ie the watch makes it up.
  3. They’re both reasonably accurate but the power numbers required for running vs cycling are just fundamentally different - although I’d figure cycling would need more due to the higher speed and increase air resistance, but then again there’s lower ground resistance so maybe not.

Any thoughts? Anyone getting comparable numbers for both? Have there been any scientific studies exploring this?

Take this for what it’s worth ($0.00, exactly!)… my son uses a Styrd power meter on his running shoes and does ultra-marathons and stuff. Running power, in my estimation, has zero direct correlation with cycling power. I have watched his ‘running’ power for 2+ years, and his cycling power (on my Kickr and on my bikes with Quarqs). I can’t see any useful comparison between the two.

For running power, we work with whatever his Stryd numbers display.
For cycling power, we work with whatever the Kickr/Quarq display.

They might as well be apples and oranges.

BUT, and this is a giant BUT, the two numbers are still watts. I know that. And watts are watts, whether they’re measured on a bike, a sneaker, or a lightbulb. My point is that we’ve set vastly different training benchmarks between the running vs. cycling disciplines. ‘Hard’ on the bike is totally different from ‘hard’ on the run. And vice/versa.

p.s. Remember the value of my comment at the start of this post: $0.00


How does a wrist based powermeter work? You do most of those activities with your legs (running and cycling) and on the bike in particular there will be little force going through your wrist :thinking:

Heart rate should reflect power output reasonably regardless of the exercise performed. There are some factors (efficiency etc.) that offset the correlation somewhat but not by that much.

The watch doesn’t do power on the bike, I use a Quarq or Assioma pedals for that. I’ve no idea how the watch does power, but I’m sure it works it out rather than measures it.

1 Like

I don’t think the numbers can be 100% compared then (apples with oranges) but its possible, in my mind at least, that a conditioned body will be able to deliver more power (maximise use their muscles more) in one sport as opposed to another.

1 Like

Even if the cardiorespiratory capacity is the same, the muscles used are different. When biking and running, white and red muscles, which use the glycogen and fatty acids around the muscles, have different usage methods. So, each sport should have different FTP and guide power.


I can’t see how this could be true. But it might be approximately true if you limit yourself to sub threshold paces over a short period of time…let’s say 90 minutes or so. For VO2max efforts heart rate will ‘saturate’ the power calculation. For efforts just barely above threshold the long component of VO2max would slowly cause power to be overestimated. For longer efforts cardiac drift would slowly introduce error as well.

I think Stryd is just measuring vertical & horizontal displacement along with respective velocity/accelerations…throwing in elevation change…and spitting out a power calculation. So if you sprint up hill you’re going to get an accurate measure at least of how many watts it took to accelerate your carcass up the hill at that acceleration and speed. Meanwhile your heart rate will just start to reflect a sprint effort somewhere at the end of your sprint.

Fun stuff.


DC Rainmaker has the most concise explanation of running power and the lack of “standard” within running let alone compared to cycling.


For running, pace is still the best training metric and just about every pro still trains to hit splits, not some power level. I nave a Stryd pod. It’s nice for measuring pace/distance in GPS limited environments. But the power readout seems like novelty to me. I still go off pace and RPE just like I have for the past 25 years.

1 Like

Yes, I did neglect to mention that the claim was only valid sub threshold. I thought that was obvious…

My point is that the heart rate won’t differ substantially at the same power output during running or cycling. If the oxygen intake is the same, so will power. The major source of error is efficiency during running compared to cycling but that is still within a couple of beats.

I think you’re onto it here. Essentially the watch doesn’t measure any power, and how would it from your wrist. It’s just a calculation / guesstimation. They call the resulting number watts and put it on the approximate same scale. But they might as well call it something else and use a scale of 0-10.

Besides, how do you like the run power?

At this point, running power is a fictional number: there isn’t even consensus on what “running power” actually means, let alone how to calculate it. And since none of the running power meters are directly measuring force production, they are blind to conditions (e.g., headwind or tailwind) that impact actual power output.

Actually , Stryd has a wind sensor and takes it into account

Neither do bike power meters, they estimate from torque, estimated from a strain gauge, multiplied by estimated cadence. That it’s all estimation is pretty clear if you look at the 4iiii gen 3 versus gen 2 reviews where you can see just how errors in the cadence estimation throw everything off. In exactly the same vein, some power meters have problems with oval cranks because it throws all the estimations off (which is why Assioma specifically points out it works for theirs). They wouldn’t care for crank shape if they were measuring direct force, but they do not.

Wrist based power estimates from pace, cadence, vertical displacement. It’s going to be less accurate than a chest mounted pod (Garmin), and even less accurate than a feet mounted one (Stryd). Likewise estimating from weather reports (Coros, Garmin) is less accurate than measuring wind on the device (Stryd).

Are the estimates good enough to train off? Sure. What’s the advantage? Maybe a small one on non-flat roads, very windy areas, for sprints and areas with bad GPS reception. (Honestly the dual band GPS removes a lot of reasons to get a Stryd IMHO, I think I’d just get a HRM-Pro if I’d start out now) Maths on power are also easier than inverse math on pace but maybe that’s just me.

Bike power and running power aren’t comparable. You don’t use the same muscles, so why would power be the same?

Except in the case of bike power meters, there is agreement for what they should measure. Yes, bike power meters are estimating power, but at least it is agreed upon on what they are estimating and they can be calibrated. Further, you can compare bike power meters against each other. You can’t compare two or more different run power meters against each other, and expect that if they were all accurate, they would agree within some error band. They are quite literally estimating different quantities.

As far as I know, Stryd and Coros should agree, and IIRC Apple’s support will be on the same scale. Garmin is an outlier.

I find this a bit of an imagined issue as I never find the need to calibrate my electric motor against my bike, and in any case there’s enough disagreement between the crank power meter and the Wahoo Kickr that I have to treat them as different powers (and I can’t really compare them!).

The existence of PowerMatch should firmly put the idea to rest that bike power meters relating to a known quantity has any meaning in practice.

In light of this, doing my Garmin power x0.75 (I don’t know if this factor is the same for everyone) and arriving at my Stryd power doesn’t feel any different practically.

So, there are some studies on this, but just as HR doesn’t have an exact relationship to cycling power, HR doesn’t have an exact relationship to running power.

TL;DR: there is a relationship, but it’s not exact. That said, there are rules of thumb that we can use.

Some running studies look at energy expenditure using the exhaled gas method: just as in cycling, you get get hooked up to a metabolic cart that measures how much CO2 you exhale. Then, since we have a pretty reasonable idea of the Krebs cycle and how energy is produced, you can calculate how much energy was expended to run at a particular speed on a treadmill. These studies show that there’s a range across individuals in their energy expenditure, but that range usually falls into something like 0.8 - 1.1 kcal/kg/km.

Thus, there’s a fairly useful “rule of thumb” which, although imprecise for any individual, is generally okay for groups of people: at submaximal running paces on flat surfaces, energy expenditure is around 1 kcal/kg/km. Importantly, this is the rule of thumb you’ll find in many exercise physiology textbooks, like McArdle, Katch, and Katch.

So, on a firm flat surface, you can use 1 kcal/kg/km. That means if you weigh 75 kg, a rule of thumb is that you’d be expending 75 kcal/km. Notice it’s not dependent on running speed (that’s a clue that this is just for aerobic running, not sprinting), so if you run a km in 5 minutes (=300 seconds), that’s 75 kcal/300 seconds, or 250 calories per second.

We know how to convert a calorie to a joule; there are about 4.18 calories per joule. But, here’s the key: human gross metabolic efficiency is around 20-25%. You may know this. If we assume GME is 23.9%, then we have another handy rule of thumb: that 1 kcal is “equivalent” to 1 kilojoule.

So work this backward, and it turns out we can get a third rule of thumb: running speed in m/s is approximately equal to watts/kg.

As a rough check, notice that with this rule of thumb, running a marathon at around 2h would be just under 6 watts/kg. So, a world-class marathoner would need to produce about the same watts/kg as a world-class cyclist. Anecdotally, people who both run and cycle say that running a 10k in 42:00 is roughly about as hard as breaking an hour for a 40k TT. Using the rule of thumb, a 42:00 10k would require about 4 watts/kg. So that’s in the right ballpark, too.

That’s for running on flat ground. We know how to calculate the energy required if you had to climb (or descend) a hill, too, so that gets added to the rule of thumb above. This is how Phil Skiba estimates running power from speed and elevation change.

Obviously, if your son is a gifted runner, his energy expenditure may be less than 1 kcal/kg/km, and if he’s more efficient than 23.9%, that will also change the exact conversion–but you can see above how and where the calculation would change.


This doesn’t make much sense to me. I am a shadow of my former self, but I can still run sub 42:00 for 10k. Nowhere near 4w/kg on the bike. Closer to 3, really. I should probably quit cycling and go back to running because clearly I am rubbish.

1 Like

This is probably correct regarding a single sided power meter but not with a dual sided. An estimate is a more or less qualified guess. In a single sided measurement one side is measured and the other side is estimated to be the same and the value is just times 2.

On the other hand, a dual sided measurement present a measured value, no guessing here, with a certain precision. This precision is usually ~±2 percent or so.