## Avg Power vs Normalised Power

colafreak
Posts: 635
Joined: Wed Dec 10, 2008 9:37 pm

### Avg Power vs Normalised Power

Hi All,

Looking at doing some TT modelling specifically where hills or headwinds are involved.

My question is, how do I keep it fair?

To explain: Obviously you can't just say, what happens if you go 2km/h faster than normal into the wind and rest when it's behind you. You can't say this because you might be assuming a power output that is impossible.

So obviously I want to model different power levels through different parts of the course.

Obviously even power over the full course is easy because we can just use a set power that's already measured. But if I say let's put more power into the headwinds and less into the tailwinds (or hills or whatever), how much more power can I assume?

Should I ensure that average power is equal to the measured 20 minute power (assuming a 20 minute TT) or should I ensure that normalised power is equal to 20MP?

Remembering of course that I must keep in line with the rest of the power curve. I.e. just because normalised and average power are below benchmark, doesn't mean I can model an output of 10 second power being put in for 1 minute without break.

Alex Simmons/RST
Expert
Posts: 4553
Joined: Fri May 02, 2008 3:51 pm
Contact:

### Re: Avg Power vs Normalised Power

Join the Google groups Wattage Forum and then view this thread and the associated paper I wrote on the subject:

I have already developed the models to both quantify optimal TT pacing and to provide strategic advice as to how best to attack a given course. I provide this service already to athletes around the world.

For TTs, using Normalised Power is a very good global constraint in my opinion, however in events such as Ironman and Half Ironman, it makes even more sense to use a global TSS constraint.

If you build your models correctly, you will find that a global NP contraint does a very good job of dealing with local maxima anyway, such that short duration power limits are generally not required as additional constraints in the modelling. One of the tricks of course is getting the model to correctly calculate NP. It's not such an easy problem to solve.

colafreak
Posts: 635
Joined: Wed Dec 10, 2008 9:37 pm

Thanks heaps.