A few weeks ago I was reading through my RSS feeds and came across a blog post that interested me.  If you’re not a powermeter freak like I am, feel free to skip over this.  It’ll bore you to death!

Here’s the jist of it:

Groover (the rider and the blog writer) did 2 rides.  The first ride (below) she described as being a tough session with the boys (although the boys were doing a recovery ride).   The average wattage was 118 watts and you can see that her heartrate (bars in red) was largely in the recovery zone.


The second ride (below) was done on the same loop but perceived as being easier, but if you look at her heartrate distribution you’ll notice that it was in much higher zones.  The average wattage was 112 watts – almost the same as the first ride, but surprisingly slightly lower considering her HR was higher.


Basically, two rides were done and had nearly the same average power, but one was perceived as being much more difficult than the other.  Let’s forget about how this was percieved for now as there are many variables that could affect this.

This is a great example of why looking at average power of a ride is not a sufficient indicator to reflect true physiological demands of the ride.   There are many variable factors that affect your rides: wind, hills, accelerations, steady tempo, descents, etc.   To be able to more accurately reflect the actual demands of a ride, Hunter Allen and Andrew Coggan developed an adjusted method to quantify the power for a ride for analysis.  They called this Normalized Power.

Normalized Power combines two factors.  First, the fact that physiological responses to rapid changes in intensity follow a time course that is predictable.  Secondly, the fact that many physiological responses (e.g. lactate production, glycogen utilization, etc) are not linear when related to exercise intensity.  For example,  a ride that has massive accelerations and then lulls (perhaps a criterium) will be more demanding on your energy systems than a ride that is much more steady – even though they may average out to being the same wattage in the end.  Normalized Power is basically an estimate of the power that you could have maintainted for the same physiological “cost” if your power output had been constant.   Because of the factors that are taken into consideration, Normalized Power gives you a better indication of the true demands of a ride than Average Power does.  This can all be found in the book “Training And Racing With A Power Meter” (I almost copied this description right out of the book).  It’s well worth the read if you have a power meter and want to learn how to use it properly.

Another good example of where average power could be misleading is in criteriums.  There is lots of coasting through corners and then hard accelerations in crits that end up giving you a low average power in the end.  However, Normalized Power will give you a better reflection of it’s true intensity.

I don’t see why the  same can’t be said if you train by using Heartrate.  If you come back from a ride and judge how intense the session was by using average heartrate, you won’t be getting a good indication of the true demands that were placed on the body.

The above power summaries from Groover’s two training sessions give a great example of how the Average Power of rides can be nearly equal, but the demands on the body greatly differ.  The Poweragent software that comes with Powertap in this case does not calculate Normalized Power, however, TrainingPeaks WKO+ software does.  I’m willing to bet that the second ride had a higher Normalized Power than the first ride if we were to analyze it on WKO+ (i.e. the ride was more difficult, even though the average power was slightly less).

I see many people who use a powermeter only look at the average power at the end of a ride to judge how difficult it was or as a basis to gauge their progression.   Hopefully the concept of Normalized Power broadens your analysis to look for more than Average Power after your ride.  Thank you Groover for letting me use your power graphs as an example to illistrate this.