Bike ·

Smart Trainer Accuracy and ERG Mode: What Your Turbo Gets Right and Wrong

Lab tests of direct-drive smart trainers find most within a few percent across normal training power, with bigger errors at the extremes. What that means for your zones, and whether ERG mode helps or hides pacing skill.

Smart Trainer Accuracy and ERG Mode: What Your Turbo Gets Right and Wrong

Sprint Summary

The short version — read this if you're short on time.

Lab tests of direct-drive smart trainers, mostly from 2016 to 2024, found them within about 1 to 2% of a reference across normal training power. One trainer had a bias of -1.1% between 250 and 700 W, but errors were bigger below 200 W and above 750 W. Precision was generally good, and one unit was still consistent after 14 months of use, though in another study two of five units of the same model shifted between tests. No study we found tested wheel-on trainers, which depend more on tyre pressure and roller contact.

ERG mode holds your power at a set number however you pedal. It makes intervals easy to hit, and indoor power varied about half as much as outdoor power in 20 male cyclists. But no study has tested whether training in ERG mode helps or harms pacing on the road, and a rowing study found that being held to a fixed pace was more physiologically demanding than self-pacing at the same average power. Use ERG for structured intervals, ride some sessions in free mode to practise pacing, and check your trainer against a second power source if you compare indoor and outdoor numbers.

Safety & Context

Indoor riding without airflow raises body temperature quickly. Use a fan, drink to thirst, and stop if you feel dizzy, confused, stop sweating or develop a headache or nausea; those can be early signs of heat illness.

Keep children and pets away from a running trainer's flywheel and the bike's drivetrain, and make sure the bike and trainer are stable before hard efforts or sprints out of the saddle.

Chest pain, unusual breathlessness, palpitations or fainting during a session need medical attention, whatever the trainer says your power is.

Full Distance

The complete research and analysis.

A smart trainer measures your power, controls the resistance and feeds an app. For many age-group triathletes it's now where most winter bike training happens. Two questions come up again and again. Is the trainer's power number accurate? And does ERG mode, where the trainer holds your power at a set value, make you a better or a lazier pacer?

This guide looks at validation studies of smart trainers, what they mean for your training zones, and the small body of evidence on fixed versus self-paced efforts. TriForward's "Why Your Indoor Watts Are Lower: Heat, Fans and the Smart Trainer" covers why many riders produce less power indoors, and "Can the Turbo Replace the Long Ride? Indoor Volume and Winter Durability" covers long indoor sessions. This piece is about the trainer itself.

As with any device study, the findings date quickly. The trainers tested below were models from 2016 to 2024, and manufacturers update hardware and firmware. Treat the numbers as a guide to the kinds of errors trainers can have, not as a ranking.

How a smart trainer measures power

There are two main types.

  • Direct-drive trainers replace the rear wheel. The bike's chain drives a cassette on the trainer, which turns a flywheel against an electromagnetic brake. The trainer works out power from the flywheel's speed and the braking force, or from a dedicated sensor.
  • Wheel-on trainers press a roller against the rear tyre. Inference: tyre pressure, roller pressure and tyre temperature all change the contact, which adds sources of error that direct-drive trainers avoid.

Both types estimate power at the back of the drivetrain, after the chain. A power meter on your cranks or pedals measures earlier, at your legs. Inference: that's one reason a trainer and a power meter can disagree by a few percent even when both are working properly.

What the studies tested

Trainer studies usually use one of two references:

  • A calibration rig, a motor that drives the trainer at a known power and cadence. This gives a true reference but no human pedalling.
  • A reference crank power meter used in many labs, with real cyclists pedalling. This includes the realities of human pedalling, but the reference has its own small error.

The numbers reported are similar to those in power meter studies: bias (average difference), limits of agreement (the range of individual differences), and the coefficient of variation or typical error (how consistent the readings are). TriForward's "Do You Need a Bike Power Meter? Accuracy, Single- vs Dual-Sided and What It Changes" explains these in more detail.

The key studies

One trainer on a calibration rig (2016)

Zadow and colleagues, in the International Journal of Sports Physiology and Performance in 2016, tested one direct-drive smart trainer on a dynamic calibration rig at 100 to 999 W and cadences of 80 to 120.

  • Between 250 and 700 W, the trainer was accurate, with a bias of -1.1% and limits of agreement from -3.6% to 1.4%.
  • At 100 to 200 W, the bias was 4.5%, with limits from -2.3% to 11.3%.
  • At 750 to 999 W, the bias was 13.0%, with very wide limits.

The authors advised caution below 200 W and above 750 W. Inference: many age-group triathletes do their endurance riding below 200 W, so the low end matters more for them than for the elite cyclists such tests often have in mind. This was a single unit, so it can't show how much units vary.

The same trainer after 14 months (2018)

The same group retested that trainer after 14 months of regular use (about an hour a week, roughly 60 hours), again on the calibration rig, at 100 to 600 W and 80 to 100 rpm.

  • The ratio bias between the two tests was 1.002, essentially no change.
  • Typical error between tests was 3.1 W, or 1.6%.

So at least one unit held its calibration well with moderate use.

Five units of the same model (2016)

Hoon and colleagues, in the Journal of Science and Cycling in 2016, tested five units of a direct-drive trainer, plus a reference crank power meter, on a calibration rig from 50 to 400 W. Each was tested twice.

  • Mean error was -1.5% for the trainers (range across units -3.1% to 0.0%) and -1.6% for the reference crank meter.
  • Typical error was 1.5% for the trainers (range 1.1 to 1.9%) versus 0.7% for the reference.
  • Two of the five trainers changed significantly between tests, with an average change of 1.3% versus 0.4% for the reference.

The authors judged accuracy similar to the reference meter, with lower but acceptable reproducibility. The key detail for owners is that units of the same model differed by up to about 3%.

Real cyclists on three units of another trainer (2018)

Lillo-Bevia and Pallarés, in the International Journal of Sports Physiology and Performance in 2018, had 11 cyclists complete 44 graded tests from 100 to 500 W, at 70, 85 and 100 rpm, seated and standing, on three units of a different direct-drive trainer, compared with a reference crank meter.

  • Agreement was very high: intraclass correlations of at least 0.996.
  • Bias was small, between -5.5 and 3.8 W in most conditions, with a few minor differences at the lowest and highest workloads.
  • Precision was high: coefficients of variation under 1.2% from 150 to 500 W.

This study included human pedalling, standing efforts and more than one unit, which makes it more relevant to real training than rig tests alone.

Older age-group riders on a newer trainer (2024)

Morais and colleagues, in the Journal of Functional Morphology and Kinesiology in 2024, compared a recent direct-drive trainer with dual-sided pedal power meters in 10 regional-level cyclists with an average age of 45.6, at power stages from 100 to 270 W.

  • There were no significant differences between the trainer and the pedals at any stage.
  • More than 80% of individual readings fell within the limits of agreement.

The sample was small and the abstract doesn't give the sex split, but the riders and power range are close to those of many age-group triathletes.

What the trainer studies add up to

Across these studies, direct-drive trainers were typically within 1 to 2% of a reference across normal training power, and precise from one effort to the next. The weak spots were very low and very high power, unit-to-unit variation, and occasional drift between tests. These are mostly rig and lab tests, at steady power, in cyclists rather than triathletes, and none tested a wheel-on trainer.

Indoor and outdoor numbers

A trainer can be accurate and you can still produce different power indoors. In a 2019 study by Jeffries and colleagues in the International Journal of Sports Physiology and Performance, 20 male competitive cyclists rode 20-minute time trials outdoors on a circuit and indoors on a lab ergometer, using the same bike and power meter.

  • Average power was 280 W indoors and 284 W outdoors, not significantly different.
  • Within-rider variability was much higher outdoors: a standard deviation of 69 W outdoors versus 33 W indoors, about twice as much.

Inference: roughly half the moment-to-moment variation in outdoor power came from the road itself, from hills, corners, wind and traffic. Indoors, you ride a smoother effort even when you're pacing yourself. TriForward's "Smooth Beats Surgy: How Bike Power Variability Shapes Your Run Split" explains why smooth power helps the run off the bike.

What ERG mode is

In ERG mode the trainer adjusts resistance so that your power stays at a set target whatever your cadence. Pedal faster and resistance drops; pedal slower and it rises. In a structured session, the app changes the target at each interval.

The appeal is obvious: you hit every interval exactly, you don't have to watch numbers, and you can't drift too hard or too easy. The concern is that you also don't practise judging effort, which is what you need on a windy, hilly race course.

What the evidence says about fixed versus self-paced efforts

No study has tested ERG-mode training against free riding on a smart trainer. The closest evidence compares fixed and self-paced efforts in single sessions, and it points in two directions.

Fixed pace felt harder in rowers (2009)

Lander, Butterly and Edwards, in the British Journal of Sports Medicine in 2009, had 9 well-trained men row 5,000 m on an ergometer twice at the same average power: once self-paced at a fixed effort rating, and once at an enforced constant pace matching that power.

  • Times and average power were the same.
  • Core temperature, blood lactate and muscle activity were higher in the enforced constant-pace trial.
  • Power varied more stroke to stroke when the rowers paced themselves.

The authors concluded that being held to a fixed pace is a greater physiological challenge than self-pacing at the same intensity, because self-pacing lets you make small adjustments as you go.

Even pace was easier in cyclists (2012)

Thomas and colleagues, in the European Journal of Applied Physiology in 2012, had 10 well-trained male cyclists ride a self-paced 20-km time trial, then matched trials at even power and at variable power with the same work and time.

  • Blood lactate was lower in the even-paced trial than in the self-paced and variable-paced trials.
  • The authors reported lower ratings of effort and more positive feelings in the even-paced trial, although the statistics in the abstract for this are unclear.

Making sense of the two

These findings aren't as contradictory as they look. In the cycling study, the even pace was set at the rider's own average from a self-paced ride, so it was a smoother version of what they had chosen. In the rowing study, the enforced pace removed the small adjustments riders normally make. Inference: holding a steady power that you've chosen and can sustain may be efficient, while being locked to a target with no freedom to ease off when fatigue builds may be harder. Both studies were small, all-male and single sessions, and neither used a smart trainer.

When ERG mode helps and when it doesn't

Built from the evidence above, and labelled as inference:

ERG mode is useful for:

  • Structured intervals at precise power, especially short ones where reacting to the number would make you overshoot.
  • Tired or busy days, when you want to do the session without managing it.
  • Threshold and tempo blocks where steady power is the whole point.

Free (level or slope) mode is useful for:

  • Practising pacing, including holding power on a simulated climb and adjusting to fatigue.
  • Sprint and all-out efforts, where ERG can't respond quickly enough and can hold you down or overshoot at the start.
  • Race simulations, because no race holds your power for you.

A common problem in ERG mode is the "spiral": if cadence drops late in a hard interval, resistance rises to keep power constant, which drops cadence further. Inference: free mode or a lower target lets you recover rather than grinding to a stop.

How to check your own trainer

  1. Do the spin-down or calibration the manufacturer recommends, after warming the trainer up.
  2. Compare with a second power source. If you have a crank or pedal power meter, ride 20 steady minutes with both recording. A consistent gap of 1 to 3% is normal and can come from where each measures. A gap that grows with power, cadence or time is worth a closer look.
  3. Check the ends of the range. Ride a few minutes at a very easy power and do a short sprint. Studies found the biggest errors there.
  4. Retest after moving the trainer, a firmware update or a long gap.
  5. Choose one source for your zones. Inference: training indoors and outdoors on two sources that differ by 3% is fine if your zones are set from the same one you train with, or you know the offset.

A triathlon scenario

The following example is illustrative, not drawn from a study. A 45-year-old preparing for a full-distance race does all his winter bike training on a direct-drive smart trainer in ERG mode. In spring, his first outdoor long ride goes badly: he surges up every hill, struggles to hold steady power on the flat, and is wrecked by the end.

  • He compares the trainer with his crank power meter. The trainer reads about 2% higher, so his outdoor zones are a little lower than he thought.
  • He keeps ERG mode for his midweek threshold intervals.
  • He moves his weekend long indoor ride to free mode on a simulated rolling course, holding a power range by feel and checking the number.
  • By the race, his outdoor rides are smoother and his power variability on hills has come down.

Common mistakes

  • Assuming trainer and power meter should match exactly. A few percent is normal.
  • Mixing zones across devices without knowing the offset.
  • Skipping calibration after moving or updating the trainer.
  • Doing all training in ERG mode and never practising pacing.
  • Using ERG for sprints, where the trainer can't keep up.
  • Riding without a fan. TriForward's "Why Your Indoor Watts Are Lower" covers why airflow matters.

What we don't know

  • Whether ERG-mode training changes pacing skill or race results. No study has tested it.
  • How wheel-on trainers compare, since we found no published validation.
  • How current models perform, since studies lag the products.
  • How accurate trainers are in long sessions, as they heat up over hours.
  • Whether women, older riders or triathletes respond differently to fixed and self-paced efforts. The pacing studies were all in trained men.

References

  • Zadow EK, Kitic CM, Wu SS, Smith ST, Fell JW. Validity of power settings of the Wahoo KICKR power trainer. International Journal of Sports Physiology and Performance, 2016;11(8):1115–1117. doi:10.1123/ijspp.2015-0733
  • Zadow EK, Kitic CM, Wu SSX, Fell JW. Reliability of power settings of the Wahoo KICKR power trainer after 60 hours of use. International Journal of Sports Physiology and Performance, 2018;13(1):119–121. doi:10.1123/ijspp.2016-0732
  • Hoon MW, Michael SW, Patton RL, Chapman PG, Areta JL. A comparison of the accuracy and reliability of the Wahoo KICKR and SRM power meter. Journal of Science and Cycling, 2016;5(3):11–15. Repository record
  • Lillo-Bevia JR, Pallarés JG. Validity and reliability of the Cycleops Hammer cycle ergometer. International Journal of Sports Physiology and Performance, 2018;13(7):853–859. doi:10.1123/ijspp.2017-0403
  • Morais JE, Bragada JA, Magalhães PM, Marinho DA. The accuracy and reliability of the power measurements of the TACX Neo 2T smart trainer and its agreement against the Garmin Vector 3 pedals. Journal of Functional Morphology and Kinesiology, 2024;9(3):138. doi:10.3390/jfmk9030138
  • Jeffries O, Waldron M, Patterson SD, Galna B. An analysis of variability in power output during indoor and outdoor cycling time trials. International Journal of Sports Physiology and Performance, 2019;14(9):1273–1279. doi:10.1123/ijspp.2018-0539
  • Lander PJ, Butterly RJ, Edwards AM. Self-paced exercise is less physically challenging than enforced constant pace exercise of the same intensity: influence of complex central metabolic control. British Journal of Sports Medicine, 2009;43(10):789–795. doi:10.1136/bjsm.2008.056085
  • Thomas K, Stone MR, Thompson KG, St Clair Gibson A, Ansley L. The effect of self- even- and variable-pacing strategies on the physiological and perceptual response to cycling. European Journal of Applied Physiology, 2012;112(8):3069–3078. doi:10.1007/s00421-011-2281-9

Frequently asked questions

How accurate is a smart trainer?

Lab tests of direct-drive trainers found errors of around 1 to 2% across normal training power, comparable with a reference power meter in some tests. Errors were larger at very low and very high power, and individual units varied. No study we found tested wheel-on trainers.

Does a smart trainer lose accuracy over time?

One direct-drive trainer retested after 14 months of regular use (about 60 hours) was still within about 1.6% of its first test. But in another study, two of five units of the same model shifted between tests, so it's worth checking yours now and then.

Is ERG mode bad for pacing skill?

Nobody has tested that directly. ERG mode holds power steady, and indoor power varied about half as much as outdoor power in one study. That makes intervals easier to hit, but it doesn't practise the judgement you need on hills and in wind, so mix in some free-riding sessions.

Why does ERG mode feel harder some days?

In a rowing study, being held to a fixed pace raised core temperature and lactate compared with self-pacing at the same average power. Inference: having no freedom to ease off may make a fixed-power effort feel harder, especially when you're tired.

The evidence behind this guide

1 graded principle, 8 sources. Open one to see the sources.

Direct-drive smart trainers typically read within 1-2% of a reference across normal training power, with larger errors at the extremes and some unit-to-unit variation; whether ERG-mode training affects pacing skill is untestedModerate evidence8 sources

In practice: Treat trainer power as accurate to within a few percent across normal training power, check it against a second source if you compare indoor and outdoor numbers, and set zones from the device you train on. Use ERG mode for structured intervals but ride some sessions in free mode to practise pacing; that advice is inference, because ERG-mode training has not been tested against free riding.

Peer-reviewed studyIndependently verified
Validity of power settings of the Wahoo KICKR power trainer ↗
Zadow EK, Kitic CM, Wu SS, Smith ST, Fell JW · 2016 · International Journal of Sports Physiology and Performance
One direct-drive smart trainer tested on a dynamic calibration rig (no human participants)

At 250-700 W and 80-120 rpm the trainer showed bias -1.1% (95% limits of agreement -3.6% to 1.4%); across 100-999 W bias was 4.2% with wide limits, driven by 4.5% bias at 100-200 W (limits -2.3% to 11.3%) and 13.0% at 750-999 W. Caution advised below 200 W and above 750 W. Single unit, 2016 model.

Peer-reviewed studyIndependently verified
Reliability of power settings of the Wahoo KICKR power trainer after 60 hours of use ↗
Zadow EK, Kitic CM, Wu SSX, Fell JW · 2018 · International Journal of Sports Physiology and Performance
One direct-drive smart trainer retested on a calibration rig after 14 months (about 1 h a week) of use

Over 100-600 W at 80-100 rpm, ratio bias between tests 14 months apart was 1.002 (limits 0.992-1.011); ICC 1.00; typical error 3.1 W (1.6%). Single unit.

Peer-reviewed studyIndependently verified
A comparison of the accuracy and reliability of the Wahoo KICKR and SRM power meter ↗
Hoon MW, Michael SW, Patton RL, Chapman PG, Areta JL · 2016 · Journal of Science and Cycling
5 Wahoo KICKR units and 1 SRM tested on a calibration rig (no human participants)

Rig-driven incremental test from 50 to 400 W. KICKR mean error -1.5% (range -3.1 to 0.0%) vs -1.6% for the SRM; typical error 1.5% vs 0.7%; two KICKR units changed significantly between tests. One trainer model, 2016. Cited by repository URL because the listed DOI did not resolve.

Peer-reviewed studyCross-sectionalIndependently verified
Validity and reliability of the Cycleops Hammer cycle ergometer ↗
Lillo-Bevia JR, Pallarés JG · 2018 · International Journal of Sports Physiology and Performance
11 cyclists (sex split not given in abstract), 44 graded tests on 3 units of one direct-drive trainer

At 100-500 W, 70-100 rpm, seated and standing, the three trainer units agreed with a reference crank power meter (ICC at least 0.996), with bias -5.5 to 3.8 W in most conditions and minor differences at the lowest and highest workloads; CV under 1.2% from 150 to 500 W.

Peer-reviewed studyCross-sectionalIndependently verified
The accuracy and reliability of the power measurements of the TACX Neo 2T smart trainer and its agreement against the Garmin Vector 3 pedals ↗
Morais JE, Bragada JA, Magalhães PM, Marinho DA · 2024 · Journal of Functional Morphology and Kinesiology
10 regional-level cyclists, mean age 45.6 ± 6.4 y (sex split not given in abstract)

A direct-drive smart trainer and dual-sided pedal power meters showed no significant differences at power stages from 100 to 270 W, good reliability, and more than 80% of Bland-Altman points within the limits of agreement. Small sample.

Peer-reviewed studyIndependently verified
An analysis of variability in power output during indoor and outdoor cycling time trials ↗
Jeffries O, Waldron M, Patterson SD, Galna B · 2019 · International Journal of Sports Physiology and Performance
20 male competitive cyclists (VO2max 60.4 mL/kg/min)

Randomised 20-minute time trials outdoors on a racing circuit and indoors on an electromagnetically braked ergometer, same bike and power meter. Mean power 280 W indoors vs 284 W outdoors, not different (P = .256), r = .94; within-person power variability far higher outdoors (SD 69 vs 33 W). All-male.

Peer-reviewed studyIndependently verified
Self-paced exercise is less physically challenging than enforced constant pace exercise of the same intensity: influence of complex central metabolic control ↗
Lander PJ, Butterly RJ, Edwards AM · 2009 · British Journal of Sports Medicine
9 healthy well-trained men, 5,000 m rowing ergometry

At matched average power (about 84% of maximal time-trial power), an enforced constant pace raised core temperature (38.72 vs 38.46 °C), post-test lactate (6.19 vs 5.24 mmol/L) and muscle activity compared with self-pacing at a fixed effort rating; performance times did not differ, and power varied more stroke to stroke when self-paced. Rowing, men only, single sessions.

Peer-reviewed studyIndependently verified
The effect of self- even- and variable-pacing strategies on the physiological and perceptual response to cycling ↗
Thomas K, Stone MR, Thompson KG, St Clair Gibson A, Ansley L · 2012 · European Journal of Applied Physiology
10 well-trained male cyclists

In time- and work-matched 20-km trials, even pacing (at the self-paced mean power) lowered blood lactate compared with self-paced and variable-paced trials; the abstract reports lower effort ratings and more positive affect with even pacing, though its reported statistics for this are unclear. Men only, single sessions.

Full write-up of this principle

Get The Forward

Research-led triathlon guidance, straight to your inbox. No spam, no filler — unsubscribe any time.