Skip to content
The ERG Room

Home » Power Zones & FTP

Normalized Power vs Average Power: What Is the Difference?

Average power tells you how much work you did; normalized power estimates how hard it felt to your body. How NP is calculated, why the two numbers differ and which one to look at.

November 20, 2026 · 4 min read

Normalized Power vs Average Power: What Is the Difference?
Photo: Engin_Akyurt / Pixabay

After a ride, most apps show two power numbers: average power and normalized power (NP). On a steady trainer session they are almost the same; after a hilly group ride or a criterium they can be far apart. Understanding normalized power vs average power helps you judge how hard a ride really was and is the key to metrics like Intensity Factor and TSS.

Average power: the simple number

Average power is the total of all your power readings divided by the time. It is a direct measure of the work you did: average watts multiplied by the ride time gives the total energy you put into the pedals. Most apps let you choose whether zeros, from coasting or stopping, count in the average. Including them gives a lower, but more honest, figure for the work done.

The problem is that average power treats every watt the same. A ride with repeated hard surges and easy stretches can have the same average as a steady ride, yet feel much harder and leave you far more tired.

Normalized power: the physiological estimate

Normalized power was developed by Dr. Andrew Coggan. TrainingPeaks describes it as an estimate of the steady power you could have held for the same physiological cost if your output had been perfectly constant. It accounts for two facts: your body responds to changes in intensity with a delay, and the strain of riding rises much faster than linearly as power increases, so hard efforts count for more.

How NP is calculated

  1. Take a 30-second rolling average of your power data.
  2. Raise each of those values to the fourth power.
  3. Average the results.
  4. Take the fourth root of that average.

The fourth power is what gives hard efforts extra weight. NP is not meaningful for very short efforts; TrainingPeaks calculates it only for rides or segments longer than 30 seconds, and it is most informative for whole rides or long intervals.

Worked examples

Steady ride. One hour in ERG mode at exactly 200 W. Average power is 200 W and NP is also 200 W, because there is no variation.

Variable ride. Thirty minutes at 300 W followed by 30 minutes at 100 W. Average power is still 200 W, but NP comes out at roughly 250 W. Your body paid for the hard half much more than the easy half saved.

Ride with coasting. Forty-five minutes at 240 W and 15 minutes freewheeling downhill at 0 W. Average power including zeros is 180 W; NP is about 223 W, closer to what the effort felt like.

These are simplified calculations that ignore the short transitions, but they show the pattern: the more variable the ride, the further NP rises above average power.

Variability index

Dividing NP by average power gives a ratio often called the variability index (VI). A value close to 1.0 means a very even ride, typical of steady trainer workouts, time trials and well-paced long rides. Higher values mean a surgy ride, such as a fast group ride, a criterium or a mountain bike race. In the variable example above, VI is about 1.25.

VI is useful for pacing. If your long rides or time trials show a high VI, you are probably surging on climbs and coasting elsewhere, which usually costs energy for the same finishing time.

When to use which number

QuestionLook at
How much work did I do? How much energy did I use?Average power
How hard was this ride for my body?Normalized power
How hard relative to my threshold?Intensity Factor (NP ÷ FTP)
How much training stress did it cause?TSS, which uses NP and duration
Did I pace evenly?Variability index (NP ÷ average)

NP feeds directly into Intensity Factor and Training Stress Score. Our guide to TSS in cycling explains how those are calculated and used to plan training load.

Normalized power on structured workouts

Every workout on this site lists TSS and IF calculated from its interval structure using the same 30-second rolling method. On a steady session such as our tempo 3x20, NP sits only slightly above average power, because only the warm-up and recoveries vary. On interval sessions such as 30/30s, the gap is much larger, which is why they feel harder than their average watts suggest.

Common misunderstandings

  • NP is not "real" power. It is a model-based estimate of physiological cost, not watts you actually produced.
  • A high NP is not automatically better. On a time trial, a NP well above average power usually means poor pacing.
  • NP depends on data quality. Dropouts and power spikes distort it, and so does an inaccurate power source.
  • Do not compare NP across very different ride lengths without considering duration. That is the job of TSS.

Questions

Why is my normalized power higher than my average power?

Because your ride varied in intensity. NP weights hard efforts more heavily, so surges, climbs and sprints raise it above the average.

Can normalized power be lower than average power?

In practice it is almost always equal to or higher than average power. On a perfectly steady ride the two are the same.

Which is better for pacing a long ride?

Aim for a ride where NP and average power are close: that means even pacing, which usually saves energy.

Is normalized power the same as weighted average power?

Several apps use similar weighted power metrics under different names. They follow the same idea, though the exact calculation can differ.

Sources

  1. TrainingPeaks: Normalized Power, Intensity Factor and Training Stress Score
  2. Zwift Insider: All about TSS (Stress Points) in Zwift

Keep reading