Normalized power and "potential"

[Iktome]

What does normalized power have to say about how hard you could go for a certain time period?

My highest normalized power for 5 minutes is about 17% higher than my highest 5 minute average power. Should I be able to produce the 17% higher number (assuming consistent power output)?

My max normalized power value came from a series of 1 minute intervals up a short hill with about 1 minute rests down the hill (at near zero power). Does this type of variable output "confuse" the normalized power algorithm? Or is 5 minutes too short for normalized power to be really useful?

Thanks for your thoughts...

 

[acoggan]

What does normalized power have to say about how hard you could go for a certain time period?

My highest normalized power for 5 minutes is about 17% higher than my highest 5 minute average power. Should I be able to produce the 17% higher number (assuming consistent power output)?

My max normalized power value came from a series of 1 minute intervals up a short hill with about 1 minute rests down the hill (at near zero power). Does this type of variable output "confuse" the normalized power algorithm? Or is 5 minutes too short for normalized power to be really useful?

Thanks for your thoughts...

Five minutes is definitely on the short end of things, and in fact if I have my way the next version of CyclingPeaks won't give you normalized power for any duration less than that. The reason for this is that the normalized power algorithm implicity assumes that whatever anaerobic work capacity you have available at the start of a period, you have that same anaerobic work capacity available when you finish. Clearly, this isn't true during all-out efforts of short duration, such that the normalized power algorithm will tend to overestimate (versus just being randomly high or low). By the time you get out to 1 h, OTOH, this effect is negligible, at which point the algorithm appears to be accurate to +/-5% or so (just my gut impression, i.e., I haven't done any formal studies...yet!).

Because of this issue, I always tell people to base their power profile on their best 5 s, 1 min, and 5 min average power...you can use normalized power to estimate the 4th value, i.e., your functional threshold power, but you can nail it down more precisely using other approaches.

 

[rmur17]

op: What does normalized power have to say about how hard you could go for a certain time period?

Aside from the restrictions Andy has mentioned, I believe NP does say a lot about how hard one could go for sustained periods of time. IIRC, NP is an estimate of the Average power one could have held for the duration if the effort were perfectly steady (as permitted by terrain, wind, group, race etc).

I track my AP versus Duration curve and that's my 'gold' standard but I also track NP versus Duration ... taking that as what I could target for a longish, flattish, well-paced effort. For example a 50 or 100m TT in the UK ....

rmur