Normalized Power (NP) vs. Average Power Calculator

This calculator demonstrates the difference between simple Average Power and Normalized Power®, a key metric used by platforms like TrainingPeaks. It helps to answer the question: does TrainingPeaks use its own algorithm to calculate average power?

NP® vs. AP® Calculator

What is the “TrainingPeaks Algorithm”?

When athletes ask, “does TrainingPeaks use its own algorithm to calculate average power?“, there’s often a misunderstanding. Average Power (AP) is a simple mathematical mean (total watts divided by time) and is not proprietary to anyone. However, TrainingPeaks heavily utilizes a specialized, trademarked algorithm called Normalized Power® (NP®).

NP® was developed by Dr. Andrew Coggan to provide a better measure of the true physiological cost of a workout than Average Power. AP can be misleading because it doesn’t account for the variable nature of cycling. A ride with many hard surges and coasting periods is physiologically much more demanding than a steady ride, even if both have the same Average Power. NP® is an algorithm that weights the more intense parts of a workout more heavily, providing a power number that better reflects the actual metabolic strain.

The Normalized Power (NP®) Formula and Explanation

The Normalized Power algorithm is not a simple average. It’s a four-step process designed to emphasize the physiological stress of power fluctuations.

1. Calculate a 30-second rolling average: First, the algorithm smooths the raw power data by calculating a moving average over 30-second windows.
2. Raise values to the 4th power: Each value from the 30-second rolling average is then raised to the fourth power. This step dramatically accentuates the higher-wattage efforts.
3. Average the results: The average of all these fourth-power values is calculated.
4. Take the 4th root: Finally, the fourth root of that average is taken. This brings the value back into the range of watts and gives you the Normalized Power.

This process ensures that high-intensity spikes, which are metabolically costly, have a much greater impact on the final NP value than periods of low power or coasting.

Variables Table

Key metrics calculated from power data.

Variable	Meaning	Unit	Typical Range
Average Power (AP)	The mathematical average of power output.	Watts	50 – 400+
Normalized Power (NP®)	An estimate of the power that could have been maintained for the same physiological “cost” if the power had been perfectly constant.	Watts	50 – 450+
Intensity Factor (IF®)	The ratio of NP to your Functional Threshold Power (FTP). It quantifies the intensity of a workout. (NP / FTP).	Ratio	0.50 – 1.15+
Variability Index (VI)	The ratio of NP to AP. It shows how smooth or “bursty” a ride was. (NP / AP).	Ratio	1.00 – 1.50+

Practical Examples

Example 1: Steady State Ride

An athlete performs a steady 30-minute effort on an indoor trainer with very little power variation.

• Inputs: Power data hovering consistently around 200 watts. FTP set to 270 watts.
• Results:
  • Average Power: ~200 watts
  • Normalized Power: ~201 watts
  • Variability Index (VI): ~1.00
• Analysis: Because the effort was very steady, AP and NP are almost identical. The VI of 1.00 confirms a smooth, non-variable ride.

Example 2: Criterium Race

The same athlete competes in a 30-minute criterium with constant hard accelerations out of corners and periods of drafting at low power.

• Inputs: Power data with many spikes to 400-600 watts and dips to 0-100 watts. FTP set to 270 watts.
• Results:
  • Average Power: ~205 watts
  • Normalized Power: ~255 watts
  • Variability Index (VI): ~1.24
• Analysis: Even though the Average Power is only slightly higher than the steady ride, the Normalized Power is significantly higher. This reflects the much greater physiological cost of the repeated surges. The high VI of 1.24 shows it was a very “bursty” effort.

How to Use This Calculator

This calculator helps you see firsthand how does trainingpeaks use its own algorithm to calculate average power by comparing it to Normalized Power.

1. Paste Your Power Data: Copy your per-second power data from a ride file and paste it into the “Power Data” text area. Ensure the values are separated by commas.
2. Set Your FTP: Enter your current Functional Threshold Power (FTP) in the designated input field. This is crucial for an accurate Intensity Factor (IF) calculation.
3. Calculate: Click the “Calculate” button.
4. Interpret the Results:
   • Normalized Power (NP®): This is the main result, showing the physiological cost of your ride.
   • Average Power (AP): Compare this to NP. The bigger the difference, the more variable your ride was.
   • Intensity Factor (IF®): This tells you how intense the ride was relative to your threshold. A value of 1.0 represents a 60-minute all-out effort.
   • Variability Index (VI): A VI of 1.00 is a perfectly steady ride. Criteriums and mountain bike races can have VIs of 1.20 or higher.

Key Factors That Affect Normalized Power

• Power Variability: The more surges and recovery periods in a ride, the higher NP will be compared to AP.
• Interval Training: Workouts with high-intensity intervals will have a much higher NP than AP.
• Coasting: Periods of zero power (coasting) lower the AP significantly but have a much smaller effect on NP due to the 30-second averaging window.
• Terrain: Hilly and rolling terrain naturally leads to higher power variability and thus a higher NP-to-AP ratio.
• Ride Duration: NP is generally considered most accurate for efforts longer than 20 minutes, as the 30-second rolling average needs sufficient data.
• Pacing Strategy: An aggressive pacing strategy with many attacks will result in a higher NP than a conservative, steady pacing strategy.

Frequently Asked Questions (FAQ)

1. So, does TrainingPeaks calculate Average Power differently?

No, TrainingPeaks calculates Average Power using the standard mathematical formula. The confusion arises because it also prominently displays Normalized Power®, which is a different, more complex calculation.

2. Is Normalized Power® a trademark?

Yes, Normalized Power®, Intensity Factor® (IF®), and Training Stress Score® (TSS®) are all registered trademarks of TrainingPeaks.

3. Why is my Normalized Power lower than my Average Power?

This is very rare and usually indicates an issue with the data or a very short selection. For any ride longer than a few minutes with any variation, NP should be equal to or higher than AP.

4. Can I use NP to set my FTP?

The NP from a maximal, near-hour-long race can be a good estimate of FTP, but it’s not the primary method for determining it. A formal 20-minute FTP test is more standard.

5. Is the algorithm the same on my Garmin/Wahoo device?

Generally, yes. Most modern head units license and use the same official NP® algorithm. However, there can sometimes be minor discrepancies due to how each device handles data recording (e.g., zero-averaging, smart recording).

6. What’s a “good” Variability Index (VI)?

It’s context-dependent. A time trial should have a VI close to 1.0 (e.g., 1.05 or less). A criterium might have a VI of 1.20 or higher. A high VI isn’t “bad”; it just describes the nature of the ride.

7. Why shouldn’t I use NP for short intervals?

The algorithm uses a 30-second rolling average. For intervals shorter than a few minutes, this smoothing process can distort the true intensity, as it takes time for the rolling average to “catch up”. For short efforts, Average Power is often a better metric.

8. Does Strava use the same algorithm?

Strava uses its own version called “Weighted Average Power,” which is similar in concept but not identical to the trademarked Normalized Power®. This can lead to different values between platforms.