CoachingAnswer

HOW ACCURATE IS STRAVA ESTIMATED POWER?

By Anthony WalshRoadman CyclingUpdated 9 July 2026

ANSWER

Strava estimated power is typically 10-30% inaccurate compared to a real power meter, and it almost always overestimates. It calculates power from speed, gradient, and rider weight using a physics model that guesses your aerodynamic drag and rolling resistance — variables it cannot measure. Tailwinds, drafting, and road surface all introduce errors it cannot correct for. Strava estimated power is useful for rough comparisons between similar rides but should never be used to set training zones, track FTP, or make training decisions.

KEY TAKEAWAYS

Strava estimated power is typically 10-30% higher than real power meter data.
It uses a physics model that guesses aero drag and rolling resistance — not measured values.
Tailwinds, drafting, and surface quality introduce errors Strava cannot detect.
Do not use estimated power for zones, FTP tracking, or training decisions.

WHO THIS IS FOR

IS THIS YOU?

The rider without a power meter

You see power numbers on Strava and want to know if they mean anything.

The rider comparing estimated to measured power

You just bought a power meter and your numbers do not match Strava's estimates.

THE ROADMAN VIEW

Anthony's view on this is blunt: Strava estimated power is entertainment, not data. It is a physics model that takes your speed, the gradient, your weight, and makes assumptions about everything else — your aero drag, your tyre rolling resistance, the wind, whether you were sitting on someone's wheel. Those assumptions are wrong often enough that the number it produces can be off by 50 watts or more on any given ride.

The problem is not that the model is bad — the physics is sound. The problem is that the inputs it cannot measure (wind, draft position, road surface) have a massive effect on the output. A 20 km/h tailwind on a flat road makes Strava think you are producing 30-40% more power than you actually are. Sitting in a group at 40 km/h requires half the power of riding solo at the same speed, but Strava assumes you are solo.

Use it for rough ride-to-ride comparisons on the same route in similar conditions — is this week's estimate higher or lower than last week's on the same loop? That trend is loosely informative. But the moment you start setting zones from it, tracking FTP from it, or comparing your estimated power to someone else's measured power, you are making decisions from bad data.

EXPERT EVIDENCE

WHAT THE EXPERTS SAY

Power measurement — Roadman podcastAs discussed across cycling tech episodes
Estimated power has a place as a rough proxy when no power meter is available, but it should never replace measured data for training decisions. The error margin is too large and too inconsistent to build a training plan on.

PRACTICAL APPLICATION

DO THIS WEEK

Treat estimated power as a rough guide only
Use it to compare similar rides on the same route in similar conditions. Do not compare estimated power to someone else's measured power — the numbers are not equivalent.
Do not set FTP from Strava estimates
If you want to train with power zones, invest in a power meter or use a smart trainer that measures power directly. Setting zones from overestimated power means every interval is too hard and every easy ride is too easy.
Weight and gradient are the only accurate inputs
Strava knows your weight (if you enter it honestly) and the gradient (from GPS elevation). Everything else — drag, rolling resistance, wind — is an assumption. On climbs where speed is low and gradient dominates, estimated power is less wrong. On flat roads where aero drag dominates, it is most wrong.

COMMON MISTAKES

WHAT CYCLISTS GET WRONG

MISTAKETreating Strava estimated power as equivalent to measured power.
FIXIt is not. A 10-30% systematic overestimate makes direct comparisons misleading. Always check whether a number is estimated or measured.
MISTAKESetting FTP or zones from estimated power.
FIXThis guarantees your zones are wrong. Use a power meter, smart trainer, or heart-rate-based zones instead.
MISTAKEComparing your estimated power to another rider's measured power.
FIXYou are comparing a guess to a measurement. The estimated number is almost always higher, which makes the comparison meaningless.

FAQ

FREQUENTLY ASKED QUESTIONS

Why does Strava overestimate power?

The model assumes you are riding solo into still air with standard tyres on a smooth road. Any tailwind, drafting, or smooth surface makes you faster than the model predicts at a given power, so it overestimates the power required to produce that speed.

Is Strava estimated power accurate on climbs?

More accurate than on flats, because on steep climbs, gradient and weight dominate the calculation and aero drag matters less. On a sustained 8% gradient, estimated power might be within 5-10% of measured power. On a flat road, the error can exceed 30%.

Should I buy a power meter instead of relying on Strava?

If you are training seriously and using power zones, yes. A power meter gives you accurate, ride-specific data that Strava estimates cannot match. If you ride casually without zones, Strava estimates are fine for rough comparisons.

How does Strava calculate estimated power?

It uses a physics model: total resistive force (gravity, aero drag, rolling resistance) multiplied by speed equals power. It measures speed and gradient from GPS, uses your entered weight, and assumes standard values for drag coefficient, frontal area, and rolling resistance.

Can I make Strava estimates more accurate?

Enter your correct weight and bike weight. Beyond that, the main error sources (wind, drafting, road surface) are things Strava cannot account for. Riding solo on a known route in calm conditions gives the most consistent estimates.

What about Garmin or Wahoo estimated power?

Similar limitations. Any estimate derived from speed and gradient shares the same fundamental problem: it cannot measure wind, draft, or surface. Some platforms use slightly different models but the accuracy range is comparable.