How Accurate Are Power Meters?

Data Summary

This was a very well-designed study with lots of careful thought put into the research plan and ensuring a good setup. What did they find with these four experiments? A lot, but I'll try to summarize here and then interpret them down below:

  • SRM variability in accuracy ranged from -10.4% to +1.0%; PT variability ranged a much tighter -2.0 to -2.9%.
  • Once corrected to a range of 0.0-2.1% accuracy, 14/15 SRMs retained calibration within that range over the course of an 11-month racing season.
  • PTs were only tested after two days, but accuracy remained similar to the results of the first trial.
  • Some variability was observed with cadence for both SRM and PT over the tested range of 60-120 rpm. This was also exacerbated by the shifting of different gears.
  • Temperature definitely affected accuracy of the readings in both SRM and PT. This was especially true when the units were exposed to and calibrated in cold (6 C) temperatures, then tested again at warm (21 C) conditions. This suggests that the zero offsets be reset over the course of a ride involving large temperature shifts.
  • SRM seemed susceptible to sudden shifts in gearing and power outputs, likely caused by hysteresis of the strain gauges. This has implications when recording sudden bursts of power, as compared to a "steady" effort like a time trial.

More: Improve Your Power

So Should I Buy One?

Only you, your significant other, coach, and possibly your accountant can determine that. However, if you do spring for one, here are some general considerations from this study that you may want to keep in mind:

  • Most of us will not have access to high-end calibration facilities like the Aussies developed. Therefore, it may be hit and miss concerning the actual accuracy of your SRM or PT. However, at least in the case of the SRM, it seems that long-term reliability is excellent. As I noted above, this is the most important consideration with power monitors and other testing tools.
  • Wherever relevant, tests done with power monitors should ideally be done in the same gear and cadences to minimize variability.
  • Where there will be large shifts in temperature over the course of a ride, variability may occur in recorded power.
  • Technology is fast-moving in the competitive bike industry. This study was published in 2004 and submitted for publication in Sept. 2003. This meant that the study was done with old, 2000-2002 era SRMs (serial numbers ranging from 235-1031) and 2002-2003 era PowerTaps. As noted by the authors, many improvements may have been made in that time to the technology in these two systems to minimize the issues reported here.
  • References
    1. Gardner AS, Stephens S, Martin DT, Lawton E, Lee H, and Jenkins D. Accuracy of SRM and Power Tap power monitoring systems for bicycling. Med Sci Sports Exerc 36: 1252-1258, 2004.

    More: Why Power Training Requires Adequate Rest

    Active logoPowered Up? Search for a cycling event.

Stephen Cheung is an associate professor of kinesiology at Dalhousie University, with a research specialty in the effects of thermal stress on human physiology and performance. Stephen's company, Podium Performance, also provides elite sport science and training support to provincial and national-level athletes in a number of sports. He can be reached for comments or coaching inquiries at
  • 2
  • of
  • 2

Discuss This Article