Your indoor FTP is 280 watts. You go ride outside with the same number on your bike computer and blow up at the 35-minute mark.

The trainer wasn't lying. The number is just not portable.

This is the single most-misunderstood piece of how cycling power meters work. The popular cycling-internet myth — smart trainers overread by 5–7% so add wattage to your outdoor FTP — is wrong. Current-generation consumer trainers are remarkably accurate. They just produce a number that means something specific to that trainer, that drivetrain, and that mode — and the number doesn't carry over cleanly to a different measurement environment.

If you train indoors and race outdoors, this is your problem. Here's what the actual data says, where the gap really comes from, and the simple consistency rule that fixes it.

Section 01What the data actually says about trainer accuracy

When Zadow and colleagues tested the Wahoo KICKR against the SRM gold standard — the laboratory power meter against which everything else is calibrated — the KICKR showed a mean bias of just -1.1% across power outputs of 250–700 W (Zadow et al. 2016). The 95% limits of agreement were -3.6% to +1.4%. That's an extremely tight match.

The follow-up reliability study, after 60 hours of accumulated use, found the KICKR maintained acceptable reliability for 100–600 W at 80–100 rpm with low typical error (Zadow et al. 2018). It doesn't drift meaningfully with use.

The Garmin Vector pedals against SRM in laboratory conditions: no statistically significant difference in power measurement (p = 0.245). In field conditions: also no significant difference (p = 0.312) (Nimmerichter et al. 2017).

The big-picture data: Bouillod et al. (2022) reviewed 74 power-meter validation studies in a systematic scoping review and confirmed that modern consumer power meters and direct-drive smart trainers, properly calibrated, measure within 1–3% of SRM under controlled lab conditions.

So where does the "my outdoor power is 25 W lower than my indoor power" experience come from? Not from the trainer being inaccurate. From three other places.

Section 02Where the gap really comes from

One — drivetrain efficiency losses. A smart trainer measures power at the trainer (driver side, after the chain and cassette). A crank-side meter measures power at the crank (driver side, before the chain). A pedal meter measures at the pedal (before the crank). The difference between crank measurement and trainer measurement is the energy lost to drivetrain friction, typically 2–3% of total power output. So your KICKR will read about 2–3% lower than your pedal-based power meter on the same effort — not because the KICKR is wrong, but because that 2–3% turned into heat in your chain.

Two — ERG mode artifacts. In ERG mode, the trainer enforces a target wattage regardless of cadence. If you slow down, resistance increases; if you spin faster, resistance drops. This hides real-world inconsistency. Outside, holding 280 W requires you to manage gradient, gearing, cadence, wind, and pacing — the wattage fluctuates by ±10–20 W around your target average. Indoors in ERG, the wattage is rock-stable. The cardiovascular cost of holding steady 280 W indoors is materially lower than holding averaged 280 W outdoors. Same number, different physiology.

Three — cross-device variability. The same rider testing across different devices (pedal-based, crank-based, trainer-based) can see substantial cross-device variability in real-world conditions — the Bouillod systematic review documents wide ranges of agreement between consumer power meters when used outside controlled lab conditions (Bouillod et al. 2022). The devices are each accurate against SRM in lab tests, but they measure subtly different things — left-leg only vs total, instantaneous vs averaged, with vs without zero-offset — and real-world cycling exposes those differences.

None of these is the trainer being broken. All of them mean the number isn't portable.

Section 03The honest test

If you want to know how big your specific gap is, do this:

• Indoor: 20-minute steady effort on your trainer in ERG mode at what you currently call FTP.
• Outdoor: same week, same fitness, 20-minute steady climb (no descents to coast through), holding what feels like the same effort. Record the average power.
• Compare the two averages.

For most amateur cyclists with a current-generation direct-drive trainer (KICKR, Neo, Direto) plus a power-meter pedal: outdoor reads 3–6% lower than indoor for matched perceived effort. That's the drivetrain + ERG-mode gap, not an error.

Document that number for your setup. It's your portable correction factor. Apply it the way a sailor applies magnetic variation — every time, automatically.

Section 04The fix isn't accuracy. It's consistency.

The right reaction to all of this isn't to distrust your trainer. It's to be deliberate about which device is the source of truth for your training metrics.

Pick one device as your reference. If you have a power-meter pedal and a smart trainer, pick whichever you'll use more often as the canonical measurement. Train to that device's scale. Calibrate it consistently (spindown for trainers, zero-offset for pedals) before each session.

Stop comparing watts across friends or platforms. When a friend says I held 290 W for an hour, the right response isn't I should aim for 295 — it's cool, I have no way of comparing that to my own number. Their 290 W and your 290 W are not the same physical measurement.

For race-day pacing, trust the device you train with. If 95% of your training is on a smart trainer, race off feel and heart rate, not your outdoor power meter's reading — it'll under-read vs what you're used to and you'll over-pace. If you train outdoors, calibrate the outdoor meter and trust it; the trainer's higher reading is irrelevant to race pacing.

For interval prescriptions, ignore absolute wattage. The Bouillod systematic review is clear on this: cross-device comparison is unreliable enough that any training plan citing a specific wattage ("hold 240 W for 4 × 8 min") has to be interpreted against the device you measured FTP on. Convert plans into %FTP and apply that percentage to your own number, not the plan's absolute watts.

Section 05What this means for HELIOS users

HELIOS doesn't measure power. The ring tracks the four signals that don't drift between devices: heart rate, HRV, body temperature, and recovery state. These read the same on the trainer, on the road, and in the lab.

When ULTRA's daily call says hold under 145 bpm or RPE 7 today, those signals are device-portable in a way that power isn't. The ring isn't replacing your power meter — it's anchoring you to the signal that doesn't lie about the difference between indoor and outdoor work.

Section 06Bottom line

Your smart trainer isn't lying about FTP. It's reading the number accurately for the conditions it measures — trainer-side power, in ERG mode, with your specific drivetrain losses absorbed.

The gap between indoor and outdoor FTP is real, but it isn't device error. It's the sum of drivetrain losses (~2–3%), ERG-mode steadiness (the trainer hides real-world variability), and the natural difference between measured and useful power in real cycling.

The correct response is consistency. Pick one device as your reference. Apply your personal correction factor when crossing between indoor and outdoor. Convert other people's wattage targets into %FTP before borrowing them.

And retire the trainer is lying myth. It isn't. You just have to know what it's actually measuring.