Power meters have undoubtedly transformed the landscape of cycling, changing not only how cyclists train but also how they strategize in races. For many recreational and online cyclists, the proliferation of power meters has opened up a quantifiable metric for performance. However, when training with a power meter, an obvious issue arises: the power data may be very inaccurate. In such cases, how reliable is your power data?

Even power meters produced by the same factory under the same conditions can have slight variations in actual use. This is one of the first confusions that arise when discussing power meter accuracy. All power meters display a nominal accuracy range, with the best on the market claiming an accuracy of ±1%. This can easily be misinterpreted as referring to "true power," but in reality, this accuracy is relative to the power meter itself. This means that the readings of the power meter may vary by ±1% from day to day or month to month. Thus, this highlights the main flaw of power meters: their accuracy varies more concerning their own readings than regarding true power.

What is "True Power"?

"True power" is the output power calculated based on energy expenditure. As cyclists, the power unit we focus on is watts. A watt is a unit of work or energy done over time. Another way to express watts is in joules per second, where a joule is a unit of energy defined as the work done by a force of one newton acting over a distance of one meter.

On a bike, power is calculated by multiplying the force applied to the pedals by the angular velocity of the crank. In simple terms, power is the force you apply to the pedals multiplied by your cadence.

In reality, accurately measuring a cyclist's power output on a bike is much more complex than it seems, with many variables affecting the final reading. From the location of power measurement to the equipment used and environmental conditions during riding, all can impact the final power reading. Just from the power meter's perspective, the strain gauges inside and the software used to calculate power readings differ, ultimately affecting the accuracy of the power meter regarding "true power."

Consistency is More Important than Accuracy

When it comes to the reliability of training data, consistency is far more important than accuracy. If you use a power meter as a training tool rather than a comparison tool, the actual power numbers it calculates are largely irrelevant. What matters more is that the power meter provides consistent data measurements for the same output each day.

For instance, if a power meter measures 25 watts higher than actual power, it may be extremely inaccurate in terms of "true power," but as long as it consistently overestimates by 25 watts, its data is valuable as a reference. In training, having consistent data for evaluation is more crucial than obtaining an accurate number.

When you wish to compare data between multiple power meters, issues may arise. This is often encountered by cyclists when riding outdoors and using a trainer. If you notice a significant difference between the readings of your bike’s power meter and your trainer, you can run both devices simultaneously during indoor training to quickly verify or rule out your suspicions.

Weather Can Significantly Affect Some Power Meters

Some power meters come equipped with Active Temperature Compensation (ATC) designed to enhance accuracy during rides. Since power meters rely on strain gauges to measure slight deformations of materials, temperature changes can cause the strain gauges or the materials they are mounted on to expand or contract. If temperature changes are not accounted for during a ride, the accuracy of the power meter could significantly deviate from its nominal range.

For power meters with ATC, this phenomenon is less of a concern. As long as you perform a zero calibration before the ride, the power meter will handle the rest.

If your power meter lacks ATC, you need to take additional steps to maintain relatively accurate data. Besides performing a zero calibration at the start of the ride, you may need to conduct zero calibrations mid-ride during temperature changes to ensure data accuracy.

The Location of Power Measurement

Another factor affecting power meter accuracy is the location of power measurement. The easiest place on a bike to consistently measure power is the rear hub, which experiences minimal force and provides a stable measurement location. Although hub-based power meters were once very popular, they have become less common in recent years, with crank and pedal-based power meters gaining popularity in the market.

Crank or chainring-based power meters are typically more accurate and stable, as they are subjected to fewer random forces and are less likely to be bumped, damaged, or worn. Generally, OEM power meters are more accurate than those modified from standard cranks.

Pedal-based power meters are perhaps the most challenging to accurately and consistently measure power, partly due to their position, size limitations, and the irregular forces passing through the pedals. The primary advantage of pedal-based power meters is their ease of switching between multiple bikes, making them ideal for cyclists with several bikes.

The location of power measurement can significantly affect the final power reading. Hub-based power meters consider transmission efficiency losses that occur between the cyclist’s power output and the hub. However, pedal or crank-based systems do not account for this. Although transmission efficiency losses are usually less than 10 watts, this can still raise the power meter's accuracy from ±1% to ±5%.

Single-Sided or Dual-Sided

This might be the most obvious area that can lead to data errors. Compared to dual-sided or total system power meters that measure the rider’s total output, single-sided power meters will always leave a margin of error. No rider has perfect left-right symmetry, and while some riders have minimal left-right leg differences, all riders will exhibit discrepancies in power output between their legs. This difference is almost negligible at low intensities but becomes more pronounced as the rider approaches maximum power output. This means that for those aiming to measure short bursts of power accurately, single-sided power meters may significantly underestimate or overestimate "true" power.

For example, at a peak power of 1000 watts, if your left-right balance is 48% to 52%, a left-sided power meter will read 48% of your output, which is 480 watts, and when doubled, yields a calculated reading of 960 watts. Conversely, if using a right-sided power meter, it will read 52% of your output, which is 520 watts, and when doubled, gives a reading of 1040 watts. In this simple example, a 4% left-right power imbalance could lead to an 8% difference in power readings.

Thus, single-sided power meters are best suited as entry-level tools for power training or for more general data collection, rather than for specific training. Muscle imbalances can also vary over time, meaning the percentage difference in output between each leg can change, making manual compensation less accurate.

Calibration and Zero Calibration

Another source of confusion lies in the distinction between calibration and zero calibration. Sometimes zero calibration is referred to simply as calibration, which further exacerbates the confusion. Zero calibration essentially "zeros out" the power meter for the ride, similar to using the tare button on a kitchen scale, temporarily defining the "no load" state of the pedals, crank, chainring, or hub.

Calibration is a more complex process usually performed by the manufacturer in a factory setting. It defines the power meter's "slope," or the definition of watts. Over time, a power meter may deviate between "true power" and its readings, necessitating recalibration to reset the power meter's slope. Specifically, this typically involves suspending a known mass from a specific point on the power meter, a process generally not performed by individual users and requires factory recalibration.

Is Power a Reliable Metric?

Power can be a very reliable metric, but it is not infallible. If you are looking for an absolute training metric, a power meter may still have some shortcomings, especially when you want to compare training data between different bikes or between indoor and outdoor rides.

While power is one of the most objective ways to collect training data, the validity of the data primarily depends on comparisons made with other data collected from the same power meter. If you want to compare your power files with those of friends, the specific numbers are mostly irrelevant, as the accuracy of measured true power may vary.

When cyclists attempt to assess their current performance against historical performances recorded with different power meters, accurately comparing cycling abilities can become challenging if you switch to different equipment after a training hiatus. For those unaware of this phenomenon, it can be very disheartening.

Heart Rate Data Still Holds Great Value

Heart rate data remains highly valuable in this era. Factors such as sleep quality, nutrition, stress, and training load all influence a cyclist's power output on the bike each day, which is where heart rate data serves as a useful reference.

Combining power output with heart rate data provides a more comprehensive view of athletic performance. If you ride at the same power output on different days but experience significant variations in heart rate, this indicates that your internal effort to produce that power is also different.

If you feel fatigued, dehydrated, or overly stressed, your heart rate will respond accordingly, manifesting as a sluggish response to intensity changes and maintaining higher than usual readings. While this knowledge does not alter your actual output, it provides more context for power output and could be helpful in data-driven training adjustments.