Powermeter data
 

Hi, my name is Laura and I`m doing a small research on powermeter software to study their characteristics. I need some training data and I was wondering if anyone could help me , giving me some of their training data (for example a few weeks data).
Obviously if you`re interested I can tell you what I can notice from different software.

Hi Laura welcome to Bike Forums! Are you a cyclist yourself? Just asking what others may be wondering....How are you related to the company? How are you studying their characteristics? What characteristics do you speak of? What are you going to do with the information? Will this be published anywhere? Will the person get a full print out of the results?

Hi, yes I`m a cyclist myself and I belong to a small group called qed which is studying the biomechanics of cycling and will implement it in a software soon. One of the aspects we wish to examen is the relation between heart rate and power.
We will pubblish all results in our website and all data from cyclist will be kept anonimous. In addition we wish to exchange information in both directions and give all reports to individuals interested in them.

The pros tend to hide that data, but some riders post both HR and power on Strava. Look around a bit.

I have about 475 rides with both power and HR, recorded over the past 18 months or so. I dunno how useful any of it would be, as in my experience HR and power are completely unrelated, but I'd be more than willing to provide it. You know, for science.

You can PM me and I'll send a link of some. Who you are testing has a whole lot to do with what you will see. I think you will need a huge data set or rec (that wear HRMs) and use power to elite. Then there are quite a few elite that do not record power, likely more that do.
At lower effort levels there is just a lot of noise. The rider is moving, talking etc. and using HR/energy on many things other than power. Same in many races. Holding consistent power [should not be]/is not the goal. A tourest tourist might show steady effort - if they also wear a HRM and use a PM.

I think Time Trial (race) data is going to be cleaner data. Here you have a rider trying to maximize their efficiency, so all the HR beats are designed to go fast. A good rider is faster in part because they can control where their effort goes. In general these rider train for that and some use HR and some use power, and some just feel to get their optimum performance. But it is the least "noise".

This is a TT that is a 300' / 7 mile climb out and back. Winds are pretty consistent from year to year.

Match up the VOS Stage Race TT with what you see on Strava. White Mountain Road Club ? Celebrating over 35 years (1980-2017)
It is work. I have done it for the riders I care about.

https://www.strava.com/segments/3438723
The riders with the lightning bolt were using a power meter. As mentioned many do not show both HRM and power (on purpose). This is an early season race for most, so the rides here will be a bit off of peak (other the most JrM 17-18).

The popular opinion is HR can't be trusted and is variable. I think you will find that HR is very stable for elite riders. To do this they are consistent about their rest, hydration and every other variable they can control. Holding within 1-2 beats is not unusual.

That is what I meant by reducing noise.

Attachment 593370

Sorry to be "41ish", but sheesh, I'm not sure I'm giving data to someone who can't even spell basic words and use correct grammar!

that is a bit harh. I meen the 4um nds to be a casual recrationa place we don't has to wrry so mush about how we say it.

Don’t worry, you’re not giving up your TdF chances by letting some random millennial know that your FTP is 200 and max power is 1100.

Believe it or not. There are those in this world that grew up not speaking English. I know, it's weird.

Wut??? That's un Americun!!!! Varsa Goda...

I can't say I've met anyone into Road Cycling that didn't speak some English.

Thanks, I don`t think I can PM directly (since I haven`t reached 10 posts).

They obviously cannot be completely unrelated, because you know, human physiology. HR lags behind effort/pwr on the micro scale, but it doesn’t take long to catch up. Perhaps you’re looking on too small a time scale, and need to look for the correspondence more broadly. In my experience, maybe the lag is around 15 - 30secs on the increase, and maybe a bit longer on the decrease, depending on intensity and duration of the preceding effort.

Of course, CycleOps wrote the book on the correlation of HR to PWR, commodifying it in the PowerCal. Utilizing, amongst other resources, a huge base of rider data uploaded to their Training Camp analysis software, I think there is not much to be gained by looking at raw data gleaned from a tiny study (unless you want to create your own Powercal-like device), because it has been done on a grand scale. The big question is how can understanding the HR/PWR relationship be utilized to benefit, either for a rider or a business? I’d be curious to hear the OP’s answer.

If you could somehow magically eliminate a huge chunk of variables affecting HR, you would still have only a ballpark guess of the raw/weighted power. HR is affected by everything, power is power. A recurring trend in my data is that as the ride length increases, power output dips as HR climbs, due to fatigue/dehydration... usually. So I can assume that past hour 5 or 6, HR will be slightly elevated at a similar level of effort versus hour 1 or 2-- but looking at the HR graph I have no idea whatsoever of what the power output would be.

And that's if you're trying to predict for one person-- forget it if you're looking at multiple people. The variables are almost infinite. On the same ride on the same course over the same time at the same moving speed, my wife's avg. HR might be 30bpm higher than mine-- I'm 8 inches taller, 30% heavier, make more power, my bike is 10lbs heavier, etc. Her not having a PM, I could make absolutely no guess at all as to her power output based on her HR (any online power calculator would give a better ballpark guess.) Her resting HR is ~15bpm lower than mine. But she also gets wicked nervous before rides and in heavy traffic-- I can see the HR spikes on the ride analysis.

TL;DR: power and HR are unrelated because you cannot determine one from the other.

HR is affected by hydration. By temperature. Even by technique (cadence).

Further, the (lagged) relationship between HR and power is not even relatively constant within a single workout... for example: "HR DRIFT" is a measure how much the relationship between HR and Power drifts during a session, like the 3rd differential.

It honestly sounds like the OP has no sufficiently researched the subject, and to he brutally honest, there is a lot of knowledge gained through experience that makes the research kinda come together and make sense.

Doing research into this field without also being active in the field itself, either as an athlete or a coach... IMO the odds of the project yielding anything useful are rather low.

I think you are missing the forest for the trees here, if for no other reason than Powercal does exactly that (i.e. calculate power from HR). Well, that and the fact that HR responds to activity effort level because of human physiology. That’s just how we work.

Probably what’s trpping you up is that you’re thinking X HR = Y watts, which is not the case, and subsequently, neither the basis for how PowerCal works. Rather, it’s the rate of change of HR that drives the Powercal algorithm. This makes short duration efforts difficult for PowerCal to handle, but from about 40 secs out, it does better, and total ride averages have proven to be pretty similar to direct force meters.

In fact, Powercal doesn’t even have a calibration requirement because it doesn’t predictably improve accuracy, which gets to my point to the OP, namely that it takes heaps of power and HR data to overcome the intrinsic inaccuracy of deriving power from HR. It has been done, by Cycleops, the top power meter company in the world, but it seems far fetched that your small company could get anywhere near the level and quality of analysis they have done.

So yeah, it’s just observably absurd to claim that HR is not related to power, and there is at least one commercial product from the preeminent power meter company in the world to prove it.

The real question is how much accuracy do you need?

PowerCal is a great example, I think. DCR claims it gets it in the right ballpark for average power over the course of a ride. But in terms of your power at any given minute, or even your maximum 20-minute average, it can be a few hundred percent off.

Sure, that’s the real question with regards to deciding between a PowerCal and a direct force meter, but that’s not the issue at hand.

I guess no one but Laura knows for sure what the issue at hand is, but she's part of a group that's "studying the biomechanics of cycling and will implement it in a software soon. One of the aspects we wish to examen is the relation between heart rate and power."

It's hard to imagine that accuracy (of that relationship and any inferences you draw about it) won't be an issue for that, but who knows?

You brought PowerCal up to demonstrate that it's possible to predict power from HR, but you can predict tomorrow's lottery numbers from HR, too, if you don't care about accuracy.

No, I brought up PowerCal to demonstrate HR and power are related. We don’t know what the OP is thinking to do with the data informing that relaionship, so the questions of accuracy (presumably you mean in terms of HR correlating to PWR) are of irrelevant at this point. What Powercal has proven, however, is that it’s fairly accurate at calculating average power over the duration of a workout in particular and quite typical circumstances, even more so in controlled workout scenarios such as stationary training.

But again, the pros and cons of PowerCal are not the issue at hand.

Of course they're related. If power is greater than zero, heart rate must be greater than zero as well.

Again, I think the central question in any PWR:HR analysis is how much accuracy do you need?

Well, that's the minimum you could say, isn't it, and so yes, I agree. DrIsotope does not; he said flatly they were unrelated, and doubled down on the assertion. That's why I brought up Powercal; we were not doing "PWR:HR analysis," just talking about whether it was possible, the answer to which you nailed: "of course."

They aren't isomorphic, but they are related. There are other factors in the equation, so knowing one won't give you the other, but they are both in the equation.