Normalized Power Estimate Calculator

Paste your power samples in watts to estimate normalized power.

How it is calculated

Paste power readings in watts, separated by commas, spaces or line breaks. The tool takes the fourth-power mean: NP = (Σ p⁴ ÷ n)^0.25, where p is each sample and n is how many samples there are. It also shows average power Σ p ÷ n and the variability index VI = NP ÷ average. Entries that are not numbers or are negative are silently skipped and not counted in n, so a typo disappears without a warning. The tool uses your samples as given: it does not smooth them with a rolling average first, and it does not know the time between samples, so evenly spaced readings (for example one per second) give the meaningful result. Results are rounded to whole watts.

Worked example

Six readings in watts: 150, 150, 300, 300, 200, 100. Average power: 1,200 ÷ 6 = 200 W. Fourth powers added up: 2 × 150⁴ + 2 × 300⁴ + 200⁴ + 100⁴ = 18,912,500,000. Divide by 6: 3,152,083,333. Fourth root: about 236.9. Result: 237 W normalized power against an average of 200 W, so the variability index is 237 ÷ 200 = 1.18.

Try this example →

Common mistakes

  • Pasting readings taken at uneven intervals. The tool weights each sample equally, so unevenly timed data distorts the mean; use samples at a regular spacing.
  • Including a column of other numbers, such as timestamps or heart rate, in the same paste. Every positive number is accepted as a power sample, so extra numbers change NP.
  • Expecting a typo to be flagged. A mistyped negative or text value is ignored, and the count silently drops by one.
  • Comparing a short list of samples to a whole-ride NP. With only a few values, a single surge dominates, so the figure is only an estimate.
  • Reading VI as good or bad. About 1.0 means steady power, while 1.18 as here reflects a ride with alternating hard and easy periods.

About this calculator

Normalized Power (NP), developed by exercise physiologist Dr. Andrew Coggan, is a cycling metric designed to reflect the true physiological cost of a ride with variable effort, since simple average power understates how hard surges and repeated hard efforts actually are. NP is computed with a fourth-power mean: raise each power sample to the fourth power, average those values, then take the fourth root of that average — a method that weights harder efforts disproportionately, the way the body actually experiences them.

Comparing NP to plain average power gives the Variability Index (NP ÷ average power); a value close to 1.0 means a steady, even effort, while a higher value flags a stop-and-go ride full of surges. Cyclists and coaches use NP, popularized through tools like TrainingPeaks, to gauge training load and pacing more accurately than raw average watts.

Was this helpful?

You can pick up to 3 reactions.

Comments (1)

  • Editorial team

    Two rides with the same 300 W average can feel very different. Samples of 200, 200, 200, 400, 400, 400 give an NP of 341 W and a VI of 1.14, while six samples of 300 give 300 W and 1.00. This tool takes the fourth-power mean of the values exactly as pasted and does not smooth them first, so paste 30-second averages rather than raw one-second readings. What spacing does your head unit record at?

Popular calculators

All Calculators