Are my power meter numbers accurate?

Your power meter is the anchor for FTP, training zones, and pacing. If the numbers drift, your training can miss the mark. Here’s how to judge accuracy, calibrate correctly, and keep data consistent ride to ride.

What “accurate” really means

Most modern power meters claim accuracy of ±1–2%. That’s good enough for structured training, but two points matter more than the spec sheet:

• Consistency: Do you get the same watts for the same effort on different days? Consistency drives reliable FTP and training zones.
• Location of measurement: Pedal/crank/spider meters measure before the drivetrain; hub meters measure after. Drivetrain losses (chain, cassette, bearings) typically cost 2–4%, so a hub usually reads lower than a crank or pedal at the same effort.

Small differences are normal. A steady 250 W on your pedals might show ~240 W on a hub-based trainer. What matters is knowing and controlling those differences.

Device types and what they read

• Pedal-based: Directly measures at the pedal spindle. Easy to swap between bikes, dual-sided options show left/right balance. Installation torque and correct washers matter for accuracy.
• Crank/spider-based: Measures at the crank or chainring spider. Stable and well-protected. Single-sided crank arms double one leg’s power; if you’re not 50/50, readings can be biased.
• Hub-based: Measures after the drivetrain. Typically reads a bit lower because it includes drivetrain losses. Rare now, but common on some indoor trainers.

Other technical factors to know:

• Single vs dual-sided: Single-sided meters double one leg’s power. If your balance is 54/46, total power can be off by ~8% in some efforts.
• Temperature compensation: Modern meters adjust for temperature, but large swings can still shift zero offset.
• Cadence detection: Magnet or accelerometer-based. Unusual pedaling patterns, very low cadence, or oval rings can slightly affect instantaneous readings; averages are usually fine.

Calibration and zeroing made simple

Two concepts get mixed up:

• Zero offset (user “calibration”): Tells the meter what “no torque” looks like. You should do this.
• Slope calibration: Sets how torque converts to watts. This is factory-set; don’t change it unless the manufacturer instructs you.

Zero offset: when and how

• When: Before most rides, after battery changes, after travel or re-installation, and when temperature changes by ~10°C (18°F) or more.
• How: Put the bike on the ground or a stand. Pedals installed, chain on, but no load. Unclip, stop the cranks, and follow your head unit’s “calibrate” or “zero” prompt. Don’t touch the bike until it returns a number or “success.”
• Auto-zero: Some meters auto-zero while coasting. It only works with truly zero torque; light chain tension or one foot resting on a pedal can spoil it.

Quick static torque check (optional)

If you suspect a bigger issue, a simple check can be revealing:

1. Hang a known weight from the pedal at 3 o’clock with the bike stable.
2. Enter the crank length in the meter’s settings correctly.
3. Read the raw torque (if your head unit can show it) or note the indicated power at 0 rpm if supported.

Expected torque (N·m) = mass (kg) × 9.80665 × crank length (m)

You should be within a few percent of expected. If not, contact support.

Data consistency: make your watts comparable

If you want your FTP and training zones to stay meaningful, make your setup repeatable.

• Use the same device for key tests (FTP, ramp tests) and benchmark workouts. Don’t mix devices mid-plan.
• Record at 1-second intervals. Avoid “smart recording” that can miss sprint peaks and alter averages.
• Include zeros in average power for workouts. Excluding zeros inflates averages and skews TSS/IF.
• Keep firmware updated and batteries fresh. Low batteries can cause dropouts or unstable zero offsets.
• Head unit settings: Display smoothing (3s/10s) is fine for your eyes; it doesn’t change saved data. Just be consistent.
• Bike setup: Tighten pedals/cranks to spec. Reinstalling a pedal or crank can change readings until you zero again.
• Environment: Indoors vs outdoors changes cooling and cadence patterns. Compare like for like when checking trends.

Rule of thumb: same bike, same meter, same recording settings, same type of ride. That’s how you keep FTP and zones stable.

Common accuracy questions, answered

“My trainer and pedals don’t match—who’s right?”

They may both be right within spec but measure in different places. Expect a hub or trainer to read ~2–4% lower than pedals/crank. If you must switch sources, do a steady 10–15 minute ride at ~200–250 W with both paired to separate head units, compare averages, and note the typical offset for context. Then pick one source for tests and structured work.

“My numbers drift during the ride.”

• Zero offset changed with temperature. Do a quick re-zero mid-ride after a big temperature swing.
• Mechanical changes (loose pedal/crank, chainring bolts) can add noise. Check torque to spec.
• Low battery or signal dropouts can cause spikes or dips. Replace the battery and re-pair.

“My single-sided meter seems off on climbs.”

Climbing often changes your left/right balance. A single-sided meter doubles one leg’s power, so imbalance shows up as error. Consider dual-sided, or at least keep this limitation in mind when setting FTP and comparing efforts.

“Do oval chainrings break accuracy?”

Most modern meters handle them reasonably well. You might see small differences in peak or very low cadence readings, but average power for steady efforts is typically unaffected.

A simple accuracy routine to adopt

• Before each ride: spin the cranks, stop, and perform a zero offset.
• After any reinstallation, battery change, or travel: zero offset twice, then do a short steady effort to verify normal numbers and RPE/heart rate.
• Quarterly (or if numbers look odd): do a steady-state comparison against your usual reference (e.g., indoor trainer) and note any persistent offsets.

Do this, and your watts will stay trustworthy—so your FTP, training zones, and pacing decisions remain on point.