Can You Trust Your Power Meter? 

“], “filter”: { “nextExceptions”: “img, blockquote, div”, “nextContainsExceptions”: “img, blockquote, a.btn, a.o-button”} }”>

Power meters have undoubtedly changed the landscape of performance-oriented cycling, from the way riders train to race-defining tactics. For the more recreational cyclist or virtual racer, the advent of widespread power meter use has opened up another metric for riders to flex with.

There is, however, one glaring issue with power meter use in one’s training. Power data can be wildly inaccurate. With this in mind is your power data really something to trust?

Every power meter will read ever so slightly differently, even those made in the same factory under the same conditions with have a slight variance out on the road. This is arguably where the first confusion comes from when power meter accuracy is discussed. All power meters will have a claimed accuracy displayed as a percentage but it isn’t necessarily clear what they are accurate to.

The best power meters on the market will claim an accuracy of +/-1 percent which can be mistakenly taken to mean +/-1 percent from “true power,” when in fact the accuracy is claimed against itself. In this case, it means the power meter will vary by +/-1 percent from day to day or month to month. Given this, it does highlight the main imperfection of power meters: they are accurate to themselves rather than true power.

What is true power?

The term true power is given to the calculated output based on energy expenditure. The unit of power we are concerned with as cyclists is the watt. This is a unit of work or energy over a period of time. Another denotation for a watt is one Joule per second with a Joule being a unit of energy defined as the work done by a force of one newton acting through one meter.

As far as power measurement is concerned on a bicycle the way it is calculated is by multiplying the force applied to the pedals, by the angular velocity of the cranks. In simple terms, power is how hard you press on the pedals multiplied by your cadence.

In reality, accurately measuring a rider’s power output on a bike is a harder task than it might seem with plenty of variables affecting the final reading. Everything from where the power is measured to the equipment being used and the conditions you ride in can affect the end reading. From a power meter-only perspective the strain gauges inside the meter and the software used by manufacturers to calculate a power reading differ which can ultimately affect the meter’s accuracy in regards to “true power.”

Consistency is more important than accuracy

When it comes to the reputability of your training data, consistency is far more important than accuracy. If you are using a power meter as a training tool for yourself rather than as a comparison tool the actual power figure it calculates is largely irrelevant. What matters more is that day to day the meter gives a consistent figure for the same output.

For example, a power meter can over-read consistently by 25 watts making it wildly inaccurate in relation to “true power,” however as long as it always over-reads by 25 watts the data is consistent with itself from ride to ride. When you are looking to track progress having consistent data to evaluate is far more relevant than having an accurate figure.

This does start to fall apart if you are looking to compare your data between multiple power meters which is something that a lot of riders do regularly when switching between riding outside and on the indoor trainer. If it feels like there is a large discrepancy between the reading you get on your on-bike meter and your indoor trainer meter then riding on the indoor trainer with both meters running can quickly confirm or dismiss your suspicions.

The weather can massively influence some power meters

Some power meters come with active temperature compensation (ATC) which aims to boost a power meter’s accuracy throughout a ride. As power meters rely on strain gauges to measure the microscopic deflections of the material they are mounted to any changes in temperature can cause the strain gauge, or the material it is mounted to, to expand or contact. If this is not taken into account as the temperature changes during a ride the accuracy of the power meter will start to vary way beyond its quoted limits.

For power meters with active temperature compensation this phenomenon is nothing to worry about. So long as you zero offset your power meter before the start of the ride the power meter will take care of the rest.

If you are using a power meter that does not come equipped with ATC then it does require a bit more rider input to retain comparatively accurate data. As well as carrying out a zero offset at the start of the ride it is also necessary to carry out interim zero offset calibrations as the temperature changes throughout the day.

Where is the power measured from

Something else that can affect the accuracy of a power meter is where the power is measured. The easiest place to consistently measure power on a bike is at the rear hub, this has the least input forces and is the most stable area for measurement. Although hub-based power meters were once very popular, in more recent times they have become somewhat of a rarity with pedal and crank-based power meters replacing them.

Crank or spider-based power meters are the more accurate and stable forms of power meters that are typically used. This is because they are subjected to less random forces than pedals and are protected away from knocks, damage, and wear. As a general rule power meters that were designed from the ground up are more accurate than systems that have had the meter retroactively fitted.

Pedals are arguably the hardest place to accurately and consistently measure power. This is in part due to their location, the size constraints of a pedal and also due to all of the irregular forces that go through the pedal. The main advantage of power pedals is how easy it is to move them from bike to bike making them a great option for riders with multiple bikes.

Where power is measured in the system can have a noticeable difference on the end power reading. A hub-based power meter positioned at the rear of the bike’s drivetrain will give a power reading with the drivetrain efficiency losses accounted for, and this energy loss takes place between the rider and the power reaching the hub. However for a pedal or crank-based system this is not accounted for. Although typically drivetrain efficiency losses equate to less than 10 watts this can still take a power meter’s accuracy from +/-1 percent up to +/-5 percent if compared directly with a hub-based system. This is also something to consider when using a power meter-equipped indoor trainer as this will read power at a lower value due to drivetrain frictional losses.

Single-sided or dual-sided

This is perhaps the most obvious area that can cause false data. Single-sided power meters will always leave a potentially massive margin of error when compared to a dual-sided or total system meter that measures the full output of a rider.

No rider has perfect 50/50 symmetry, some riders will have smaller differences than others but all riders will have a discrepancy between left and right leg power. At lower intensities this discrepancy can almost be negligible to the point it is insignificant however, the closer a rider gets to peak power production the more pronounced any discrepancies will become. This means that for riders looking to accurately measure short bursts of power a single-sided power meter is likely going to either under or overread significantly in comparison to the ‘true’ power produced.

If for example at a peak power of 1,000 watts, you have a left/right balance of 48-52 percent this means that a left-side-only meter would read 48 percent of your output i.e. 480 watts and double this to give a calculated reading of 960 watts. Likewise, if you have a right-side-only meter this would read 52 percent of your output i.e. 520 watts which when doubled gives a reading of 1,040 watts. In this simple example, a 4 percent left/right power imbalance equates to an 8 percent potential power reading discrepancy.

It is for this reason that single-sided power meters are best used as an introduction to training with power or for more generic data collection rather than specific training. Muscular imbalances can also change over time meaning that the percentage difference between each leg can also change making it hard to accurately manually compensate.

Calibration vs zero offset

Another area that does cause some confusion is the difference between a calibration and zero offset. This is sometimes perpetuated with a zero offset being referred to as a calibration. A zero offset is essentially zeroing the power meter for the ride much like using the tare button on a set of kitchen scales. This temporarily defines what a ‘zero’ load is to the pedal, crank, spider or hub.

Calibration is a far more involved process typically performed by the manufacturer in the factory. This is used to define the “slope” of a power meter which essentially defines what a watt is. Over time power meters can lose accuracy both in regards to true power and its own reading. This is where a calibration can be performed to reset the slope of the meter. Depending on the power meter this is typically done by suspending a known mass from a defined point on the power meter and measuring the calculated load.

Is power a reliable metric?

Power can be an incredibly reliable metric but it isn’t unassailable. If you are looking for an absolute training metric your power meter may still be something to be desired, especially if you are looking for comparable training data from bike to bike or between indoor and outdoor riding.

Although using power is the most objective way to collect training data the validity of the data is largely resigned to comparison with other data collected by the same meter. If you are looking to compare power files with friends the numbers discussed are mostly irrelevant as the variables in data can make it hard to draw any meaningful conclusions.

Where this matters more is when riders look to evaluate their current performances against historical efforts recorded with a different power meter. If you are looking to return to a certain level after a hiatus from training and are now using different equipment it can be difficult to accurately compare abilities which for those unaware of this phenomenon can be wildly demotivating.

Adding heart rate can provide a fuller picture

If power is an objective picture of the effort you are putting out then adding heart rate data gives the picture some color. As much as some might contest, cyclists are not robots and are in fact complex life forms and as such we are affected by factors such as sleep quality, fueling, stress, and training load. All of this and plenty of other factors influence a rider’s day-to-day ability to produce power on the bike.

Pairing power and heart rate together gives a far fuller representation of the effort behind the output. If you are riding at the same output on separate days with wildly different heart rates then this suggests that the internal effort to produce that power is also different.

If you are fatigued, dehydrated, or stressed your heart rate will back this up by being slow to respond to changes in intensity before sitting at a higher value than you would typically expect. Even though this knowledge will not change your actual output it does give it more context which can be beneficial in the setting of data lead training.