Interval custom Data Field appears incorrect. Normalized to Average watts

[Mitchell_Foley]

A single of many examples: *16 sec effort shows Normalized as 335w and Average as 639w?

[MedTechCD]

What custom Data Field are you talking about? And can you post a screenshot of the zoomed in interval?
Probably something is wrong because NP is (almost) always higher then Average but it is very difficult to give any meaningful reply with so little information.

[Mitchell_Foley]

This was today.
29 sec effort Normalized 420w, Average 601w

[MedTechCD]

Ah, you mean the std NP in the selection/interval fields. I thought you were referring to a custom Data Field. This has been discussed several times here on the forum.
Bottom line: NP shouldn’t be used for short intervals because you need at least 30sec of data to calculate NP. TrainingPeaks found a workaround to still use it but in reality it means that they are not respecting the definition of the metric they developed themselves…

Incorrect NP numbers

Hi there Hoping someone can point me in the right direction. For some time I’ve been suspecting unusually low NP numbers after my rides. Today this was confirmed when I was doing a seated sprint session, and when pressing the lap button after my interval, my NP for previous lap was almost half the power I was holding for the duration of the effort. I clearly recall this being way higher and very close to average power when doing this workout previously. I’m using Garmin Edge 520 and Power2ma…

Normalised Watts / Weighted Average Power

@david for the site, may I know which method of NP(R) or Weighted Average Power are you using? I noticed that you didn’t use the name NP(R) but instead used Normalised Watts (Golden Cheetah uses the term isoPower) instead. I was reading up from this link [Screenshot 2022-02-11 at 3.10.28 PM] [Picture courtesy of link above] The key difference i think is the number of samples for the rolling data. Are you using 25s or 30s? Average Cadence - Exclude Zero Average Power - Does it incl…

NP on Intervals differs from other sites

Hi, today I’ve noticed that NP of my ride on Intervals is different from NP on Connect, Trainerroad and Sauce for Strava. Intervals 281 NP, 203 W avg Connect 267 NP, 204 W avg Trainerroad 268 NP, 203 W avg Sauce for Strava 266 NP, 203 W avg As far as I know the computation should be the same since it’s the same metric. Any ideas?

[Mitchell_Foley]

Thanks! Makes sense.
So, if efforts under 30sec do not calculate accurate Normalized Power, is there an exact minimum duration that will calculate accurate Normalized Power?

[Gerald]

TrainingPeaks say you should ignore everything less than 10 minutes, so I would avoid looking for an exact number/duration. We are humans, not robots.

The formula for NP is:

1. Take the moving 30 second average of your power data.
2. Take the moving 30 second averaged power data and raise each value to the 4th power (x^4).
3. Find the average of the values from the previous step.
4. Take the 4th root ( ∜x or x^¼) of the mean value found in the previous step.

Normalized Power