Running Power Meters: What They Actually Measure and Whether a Triathlete Needs One

Most triathletes who own a bike power meter have wondered whether they need the same thing for running. Watches now display "running power" as standard, and foot pods promise watts for every stride. If power transformed how you ride, it seems obvious it should transform how you run.

The evidence is more mixed than the marketing. Running power can be useful, but it's a different kind of number from bike power, and the research behind it is small. This article explains what running power meters actually estimate, what the validity studies show, where the numbers are weakest and how a triathlete might use one.

What a bike power meter does that a run power meter can't

A bike power meter measures force directly, with strain gauges in a crank, pedal, spider or hub, and multiplies it by how fast that part is turning. The bike is a closed mechanical system. Almost all the work you do goes through the drivetrain, so a strain gauge in the right place captures it.

Running has no drivetrain. Your muscles and tendons work to move your body forwards, lift it up and down with each stride, and absorb and return energy as you land. A lot of that work is stored and released elastically in tendons, especially the Achilles, and none of it passes through a single point you can bolt a sensor to.

Running power meters get around this by estimating. A foot pod or watch uses accelerometers and gyroscopes, sometimes with a barometer and GPS, to track how your foot or body moves. A model then turns those movements, your body mass, your speed and the slope into a power figure in watts. Some devices add an estimate of air resistance.

So the question isn't "Is this measuring power?" It's "Does this estimate track the things we care about, such as how hard you're working, and is it consistent?"

What "validity" means for running power

Researchers test running power devices in three main ways:

• Against oxygen uptake (VO2). At steady submaximal efforts, oxygen uptake reflects how much energy you're using. If a device's power rises and falls with VO2, it's tracking effort.
• Against mechanical power from a force plate or force treadmill. This tests whether the watts themselves are right, at least for the mechanical work the lab can measure.
• Against itself. Repeatability, or reliability, asks whether the same run on a different day gives the same number. A device that's valid on average but noisy from day to day is hard to train with.

None of these reference methods is perfect, which is part of why the field is still arguing about what "true" running power is.

The five-device comparison

The most useful head-to-head study is by Cerezuela-Espejo and colleagues, published in the European Journal of Sport Science in 2021. They tested 12 endurance-trained men on two occasions, on both a treadmill and an outdoor athletics track. Each session included runs at different speeds, with added body weight and on different slopes, while oxygen uptake was measured with a metabolic cart.

They compared five systems: the Stryd foot pod read by its phone app and by a watch, RunScribe, Garmin Running Power and Polar's wrist-based estimate.

The results were clear-cut:

• Stryd was the most repeatable in all environments and conditions, with a standard error of measurement of 12.5 W or less, a coefficient of variation of 4.3% or less, and an intraclass correlation of at least 0.98.
• Stryd also related most closely to oxygen uptake, with correlations of at least 0.91 and a standard error of the estimate of 7.3% or less.
• The other three systems showed some relationship with oxygen uptake, but their poor repeatability led the authors to question whether they were suitable for measuring running power at all.

That's a strong result for one device. But the sample was 12 trained men, the devices were tested on the firmware available at the time, and none of the protocols lasted anywhere near the length of a triathlon run.

Tracking effort versus getting the watts right

Imbach and colleagues, writing in Sports in 2020, looked at Stryd in more detail. Six recreational runners (four men in their late thirties and two women aged around 35) ran on an outdoor track at speeds from 8 to 19 km/h. The researchers compared Stryd with a portable metabolic analyser, track-embedded force platforms and three-dimensional motion capture.

Stryd power rose closely with oxygen uptake (R² = 0.82) and with mechanical power from the force platforms (R² = 0.88). But Stryd underestimated the absolute power compared with the lab measures. The authors concluded that it gave valid measures of ground contact time and leg-spring stiffness, but underestimated absolute power output.

Taboga and colleagues found something similar with three devices in 11 runners on a force-measuring treadmill, at speeds from 2.22 to 4.44 m/s (about 8 to 16 km/h). The Stryd foot pod, a Garmin chest-strap system and a treadmill-based system all increased linearly with speed and with aerobic power, but their absolute values differed from each other. Each device had its own idea of how many watts a given speed was worth.

Taken together, these studies suggest that a good running power meter tracks changes in effort well, which is what you need for pacing. The absolute watts are best thought of as a brand-specific scale rather than a physical truth. Inference: 250 W on one device and 250 W on another may not represent the same effort, and neither can be compared directly with 250 W on your bike.

Does running power tell you about running economy?

One appealing claim is that running power reveals how efficiently you run: if two runners hold the same pace but one produces fewer watts, maybe that runner is more economical.

Austin and colleagues tested this in 17 well-trained distance runners (nine men, eight women) at lactate-threshold pace. Stryd power was positively related to running economy (r = 0.6, 90% confidence interval 0.2 to 0.8), so power explained roughly 31% of the differences in oxygen cost between runners. That's a real relationship, but a loose one. The authors concluded that the foot pod may not be accurate enough to estimate differences in running economy between competitive runners.

Inference: running power is better at telling you how your own effort changes from minute to minute than at telling you whether you're a more economical runner than your training partner.

Where running power is weakest

Hills and trails

On a flat road or track, speed and power move together closely. On hills, the relationship breaks, and that's exactly where a power reading should help. Yet slope is the condition the research covers least.

A 2020 scoping review by Jaén-Carrillo and colleagues gathered 19 studies of running power sensors. They found that slope had been tested in only a few studies, and that in one, RunScribe's error against oxygen uptake rose to 18.5% on slopes. The review concluded that more research was needed before the validity and reliability of running power meters could be firmly judged.

A small 2024 field study by Berzosa and colleagues fitted five trail runners (four men, one woman) with two Stryd and two Garmin devices each. Each brand agreed well with a second unit of the same brand. But Stryd and Garmin agreed poorly with each other, especially for power on uphill sections. The authors' advice was to stick with one device and not treat the two brands as interchangeable. With only five runners, this is a warning sign rather than a firm finding.

Wind

Air resistance matters for runners, especially fast ones or anyone running into a headwind. Some devices estimate it. The scoping review noted that Stryd's wind estimate rested on a manufacturer white paper rather than peer-reviewed testing. On a treadmill, where most studies happen, there's no wind to test. Inference: on an exposed, windy run course, power numbers are partly a model's guess at a force the research hasn't checked.

Fatigue and form changes

Running power models assume a relationship between how your body moves and the work it does. As you tire, your stride changes. The scoping review noted that how fatigue affects the power estimate is barely studied. That matters for a triathlete, because the whole run is done tired.

Power, pace or heart rate?

Each of the three main ways to gauge running effort has strengths and weaknesses.

Pace is direct and well understood. On flat ground and in calm conditions, pace is an excellent effort measure, and TriForward's "Critical Speed for Runners" shows how to set zones from it. It falls down on hills, in wind, on soft or technical surfaces, and when GPS is unreliable, such as under trees or between tall buildings.

Heart rate reflects the internal cost of what you're doing, including the effect of heat, dehydration and fatigue. That's useful, but it lags behind changes in effort by a minute or more, drifts upwards during long efforts, and is affected by caffeine, sleep and illness. TriForward's "Why the Ironman Marathon Falls Apart After 25 km" covers heart-rate drift in detail.

Power responds almost instantly, like pace, but in principle accounts for slope. That makes it potentially the best of the three for hills and for short intervals. Its weaknesses are the ones above: different absolute values between brands, limited testing on hills and in wind, and no direct information about heat or fuelling.

None of them tells you everything. Perceived effort also deserves a place. TriForward's "Racing by Feel vs. Racing by Numbers" looks at the pacing research behind combining them.

Using running power off the bike

The first kilometres of a triathlon run are where many athletes go wrong. Your legs feel strange, your heart rate is already high from the bike, and your sense of pace is unreliable. TriForward's "The Bike-to-Run 'Brick' Problem" describes why.

This is where power may genuinely help. Because it responds immediately and doesn't depend on a heart rate that's still elevated from cycling, power could give you a steadier reference than heart rate in the first 10 to 15 minutes. Inference: no study has tested whether triathletes who pace the run by power run faster than those who pace by heart rate or pace, so this is reasoning from how the devices behave, not a proven benefit.

Here's how that might work in practice, with illustrative numbers rather than data from any study. In training, you find that your steady long-brick efforts settle at around 230 W on your foot pod on flat roads. On race day, the first kilometre off the bike feels easy at 260 W, but your heart rate is already high. Rather than trusting the feeling, you hold 225–235 W. On a hill at 5 km, pace drops sharply but you keep the power in the same range. Later in the run, if heart rate climbs and you feel worse at the same watts, you slow down. The power target is a guide; heat and fatigue overrule it.

Should a triathlete buy one?

Consider a running power meter if:

• Your run course or training routes are hilly, and pace is a poor guide.
• You already train with structured intervals and want a fast-responding target.
• You'll use the same device consistently, so you can compare your own numbers over time.
• You understand the numbers are estimates, not a lab-grade measure.

It's probably not a priority if:

• You train mostly on flat roads or a track, where pace already does the job.
• You don't yet have a consistent run routine. Training consistency will matter far more than the measurement tool.
• Your budget would be better spent elsewhere, such as a bike fit, coaching or good shoes.

Many watches now offer a power reading without a separate pod. Based on the five-device comparison, those wrist-based estimates were less repeatable than Stryd at the time of testing. Newer models may perform differently, but they haven't been independently tested in the same way.

Common mistakes

• Comparing run watts with bike watts. They're different measurements made in different ways. A running FTP isn't comparable with a cycling FTP.
• Comparing numbers across brands or devices. Absolute values differ between devices, and the agreement is worst on hills.
• Changing your running form to change the number. Metrics such as "form power" aren't proven to reflect injury risk or efficiency.
• Ignoring heart rate and feel in the heat. Power can't tell you that you're overheating or running low on fuel.
• Retesting rarely. Power zones, like pace zones, need updating as your fitness changes. Re-test on the same device and under similar conditions.

What we don't know

• Whether run power improves triathlon results. No study has compared run-power pacing with pace or heart-rate pacing in triathletes.
• How accurate any device is on hills, trails and in wind. These are the conditions where power should add most, and they're the least studied.
• How the estimates hold up late in a long run, when stride and form change with fatigue.
• How well the findings apply to women and older athletes. The largest comparison included only men, and most samples were small.
• How newer watches and firmware perform. Manufacturers update their algorithms, and the published research lags behind.

References

• Cerezuela-Espejo V, Hernández-Belmonte A, Courel-Ibáñez J, Conesa-Ros E, Mora-Rodríguez R, Pallarés JG. Are we ready to measure running power? Repeatability and concurrent validity of five commercial technologies. European Journal of Sport Science, 2021;21(3):341–350. doi:10.1080/17461391.2020.1748117
• Imbach F, Candau R, Chailan R, Perrey S. Validity of the Stryd power meter in measuring running parameters at submaximal speeds. Sports, 2020;8(7):103. doi:10.3390/sports8070103
• Taboga P, Giovanelli N, Spinazzè E, Cuzzolin F, Fedele G, Zanuso S, Lazzer S. Running power: lab based vs. portable devices measurements and its relationship with aerobic power. European Journal of Sport Science, 2022;22(10):1555–1568. doi:10.1080/17461391.2021.1966104
• Austin CL, Hokanson JF, McGinnis PM, Patrick S. The relationship between running power and running economy in well-trained distance runners. Sports, 2018;6(4):142. doi:10.3390/sports6040142
• Jaén-Carrillo D, Roche-Seruendo LE, Cartón-Llorente A, Ramírez-Campillo R, García-Pinillos F. Mechanical power in endurance running: a scoping review on sensors for power output estimation during running. Sensors, 2020;20(22):6482. doi:10.3390/s20226482
• Berzosa C, Comeras-Chueca C, Bascuas PJ, Gutiérrez H, Bataller-Cervero AV. Assessing trail running biomechanics: a comparative analysis of the reliability of Stryd and Garmin RP wearable devices. Sensors, 2024;24(11):3570. doi:10.3390/s24113570