The case for Cycleop’s PowerCal
July 2012 UPDATE: I have now made a review and comparison of PowerCal. See my new post with detailed review and powertap comparison here.
One of he big, and most surprising product announcements from Cycleops two weeks ago was the all-new PowerCal heart rate monitor.
This new device is an ANT+ heart rate monitor strap with a twist… Based on your heart rate, it derives the power you are producing. This is a very surprising announcement from a company that is almost a standard and helped open up and democratize the power meters to all cycling enthusiasts.
There are other products out there that derive power from other metrics. Some do it real time like the iBike, others after the fact, taking into consideration things like weight and grade, like Strava and SportTracks, but to my knowledge this is the first one that does it from the heart rate.
It is a very smart product. It takes your heart rate, derives the power using advanced algorithms, and then broadcasts it over ANT+ so it can be read on your cycling computer. Very smart !!!
Steve Chapin, CycleOps’ marketing director said “Lab tests have shown a very consistent correlation between heart rate and power indoors, the correlation is almost one-to-one,” said Chapin. “The problem is that there’s all sorts of variables outside, hills, heat, etcetera. What we’ve done over the last few years is to analyze literally thousands of ride files and we’ve been able to identify and establish trends in instances where there’s a deviation in the correlation between heart rate and power and we’ve taken it one step further by building algorithms that identify those trends.”
The data analysis, subsequent algorithms and programming allowed CycleOps to build the PowerCal, which will sell for $199. “It’s not going to replace a PowerTap,” said Chapin, but it allows riders to have conversations about training and power who don’t have power meters.
The PowerCal must be initially calibrated for a rider using a power meter and a specific protocol in which heart rate and power information are captured then uploaded to the PowerCal unit. The unit is currently in trial at the University of Colorado to validate its accuracy. Chapin estimated the PowerCal’s accuracy to be between 5- and 10-percent based on CycleOps testing.
I remain skeptical.
Heart rate and Power generated are very different and react very differently. One of the big differences is that Power is immediate whereas Hear Rate is not.
When you accelerate and increase the power you immediately see the increase in power, but the hear rate will take a while to catch-up to the physiological change that just occurred. Every cyclist has experienced this.
The reverse is also true. When you stop powering on the pedals, power goes immediately away, whereas the heart rate will decay slowly, depending on your fitness, freshness, etc… How does Cycleops account for this?
The first I heard about this, I immediately remembered my VO2max test. There I could see a very obvious correlation between heart rate and power… That was a test under very rigorous conditions, pedaling smoothly and constantly, indoor on a trainer, very controlled environment.
The results are remarkable:
However… how does it work in the real world?
Out of curiosity, I‘ve took some of my powertap files and plotted Heart Rate vs Power, and the results do not show any type of correlation. This is caused mainly by the immediateness of power versus the inertial nature of heart rate.
The plot is presented below:
Looking at this graph, I can’t see how you’d be able to derive power from heart rate.
Even when including only cadence between 80 and 100, the diagram still does not hold much promise:
Looking at data like this, although you start to identify a trend, it’s hardly a strong correlation.
Possibly indoors, when you are always pedaling at the same cadence, it’d would work… The next graph plots watts and hear rate only when pedaling at cadence between 77 and 93 rpm.
Still no strong correlation is visible…
In conclusion… Let’s just say… I’m very curious to see side-by-side comparison, see a powertap graph against a powercal graph… Even with the calibration done with a power-heart-rate test… I simply don’t see how you can derive power from heart rate outside the lab under very strict conditions.
6/30 UPDATE: Some comments were send inquiring if results would be different if the workout would be done indoors on a trainer, instead of outside, on the road.
It so happens that I have a Computrainer, that I have for indoor training (more infrequently than I would want, but that’s a different issue). I used one of the workout files generated by the computer connected to the trainer and analyzed the heart rate vs watts:
There was in fact, a slightly better correlation between heart rate and power in an indoor environment, but still the correlation factor (0.331) is too low to formulate any conclusions.
In general, when analyzing correlation, a general way to interpret the calculated r value is as follows:
0.0 to 0.2 – Very weak to negligible correlation
0.2 to 0.4 – Weak correlation
0.4 to 0.7 – Moderate correlation
0.7 to 0.9 – Strong correlation
0.9 to 1.0 – Very strong correlation
In this case, we have a weak correlation factor, which means that the error margin is too large to generate any meaningful conclusions.
So, why is the first graph presented in this article so neat and perfect, whereas, any correlation from a workout, either outdoors or indoors, does not present any meaningful correlation?
In my opinion, this has to do with the VO2max test protocol, where you are supposed to pedal at constant cadence without any accelerations. Since the heart rate is a trailing indicator of the effort, every time there is an acceleration (or deceleration), the correlation between heart rate and power will fail. In other, it does not work for any workout where you won’t maintain a constant cadence and constant effort (or slowly increasing, which is the case of the VO2max test).
So, conclusion… I’d say that it may work in the lab (under very strict testing conditions), but does it work in real life? In my opinion, either indoor or outdoor, it will have an extremely high error of margin since the effort is not constant.
As I said, I’m very curious to see side-by-side comparison, see a powertap/computrainer graph against a powercal graph… I’m sure Cycleops will thoroughly test before launching the product, so maybe they’ll include some black magic that makes everything work.
What do you think? I’d love to hear your opinions.