What is normalised power and what can it tell us?

by Matthew de Vroet


What is the best way to measure how hard a ride is? Should you simply look at distance and time? Or do you delve deeper and analyse your average power? For those of us lucky enough to have a power meter, average power can paint part of the picture but in order to fully understand just how hard a ride is, normalised power is a much more useful tool.

So what exactly is normalised power? And why is it a more useful measure than average power?

Normalised power vs average power

Normalised power is an important metric for cyclists who train with power and it is valuable to understand how it works and when to use it. In order to build fitness, athletes put their body under stress during training. The body adapts to this stress through recovery and, in the next round of training, is then able to accommodate more stress, leading to increased athletic abilities.

Defining how “stressful” a block of training is can be a difficult task and you must consider every component, including the duration, frequency, intensity and “how hard” the rides were. While most of those aspects have straightforward answers, the “how hard” element is a difficult factor to quantify.

You can look at average power, but that is purely the average of your power throughout the whole ride. This is a simple measure to understand and at first can seem like a worthy way to measure the effort of a particular ride. However, it can be limited in its function.

For example, if you ride on a flat road for an hour at approximately 200 watts, you will average roughly that amount for your effort. This might be an easy effort and comfortable to sustain for the entire hour.

On the other hand, if you rode an undulating course and did an hour-long effort, trying to average 200 watts could be a much harder task. You would need to hold more than 200 watts uphill as the descents would neutralise those efforts and bring your average back down.

You might still manage to average 200 watts for the hour-long effort but it would be a lot harder than a flat ride at a constant pace. It will also leave you a lot more fatigued and needing more recovery.

So even though the two rides might have similar average power numbers, the energy and physical costs needed to achieve these values will vary.

This is where normalised power comes into play. Normalised power will take into account the variance between a steady workout and a fluctuating workout in order to quantify the physiological cost. This means that for a highly variable workout, normalised power will be much higher than average power. For a steady, consistent workout, the two values will be very similar.

Normalised power in action

A great example of the difference between average and normalised power can be seen in Rigoberto Urán’s data from the 2017 Tour de France, a race he finished in second overall behind Chris Froome.

Stage 20 of the Tour was an individual time trial where Urán needed a consistent power for 22.5km if he was going to improve his overall position in the race. The Colombian’s power data can be seen below:

Average power: 378W (6.00 W/kg)
Normalised power: 389W (6.16w/kg)
Difference: 3.2%

Through these figures, you can already see how Urán’s effort was extremely measured and steady. The two numbers are very similar, with the difference likely to be due to a short, two-kilometre downhill section where Urán wouldn’t have been pushing the pedals as hard. Normalised power accounts for that and adjusts accordingly.

Compare this to Urán’s data from the extremely hilly stage 9 and you can see why normalised power becomes a better metric for analysing a lumpy ride.

Average: 219W (3.78 W/kg)
Normalised: 292 W (4.63 W/kg)
Difference: 33%

The 181km stage featured three hors categorie climbs, which meant three long descents. Look at Urán’s average power for the ride and you might be tricked into thinking it was a relatively easy stage. Once you realise his normalised power was almost 80 watts higher (a difference of 33%) you can start to understand the value of the metric.

Normalised power provides a more realistic way of representing a rider’s workload than average power. It can be imagined as the power a rider could have maintained for the same physiological cost had their power been perfectly constant. The examples above show how it is particularly useful when considering hilly races where a rider’s power varies a lot.

The challenge for any athlete is to know how much stress they can put their body under and how much recovery will be required to maximise adaptation and, in turn, fitness. Measures such as normalised power allow athletes to better understand how hard a particular ride was.

Average power does serve a purpose as it gives a general overview on the ride, but for those lumpy courses, you’re better off using normalised power when trying to perceive “how hard” your effort was.

This is an updated version of an article first published on CyclingTips in 2009.

About the author

Matt de Vroet joined CyclingTips as an editorial intern in April 2017. He is a third-year journalism student at Monash University in Melbourne and currently races for Van D’am Racing in Australia’s National Road Series.

Editors Picks