Take the ramp test. That’s the short answer for 90% of first-timers. It’s quick, it doesn’t punish bad pacing, and it produces a working number you can train with starting tomorrow.
But there are three protocols, three correction factors, and one real reason the ramp test can lie to certain riders.
What FTP Actually Measures
FTP (Functional Threshold Power) is the highest average watts you can hold on the bike for roughly one hour without blowing up. Every training zone you’ll ever use flows from this number.
Think of FTP as the engine’s red line. Below it, you can sustain the effort. Cross it, and the clock starts ticking down on how long you’ll last.
No FTP test actually measures a true 60-minute effort. All three protocols estimate it using a correction factor. That’s not a flaw — it’s a practical trade-off.
The Three Protocols at a Glance
Here’s how they compare before we get into the details.
| 20-Minute Test | Ramp Test | 8-Minute Test | |
|---|---|---|---|
| Active effort | 20 min | ~15–20 min | 2 × 8 min |
| Total session time | 60–70 min | 25–30 min | 50–55 min |
| Correction factor | 20-min avg × 0.95 | Best 1-min × 0.75 | Avg of both × 0.90 |
| Pacing difficulty | High | None | Medium |
| Anaerobic bias | Moderate | High | Low |
| Best for | Experienced riders | Beginners, Zwift users | Accuracy seekers |
| Repeatability | ±15W (95% LoA) | ±5–7W typical | ±~2W in N=1 data |
The correction factors aren’t arbitrary. Coggan and Allen derived the 0.95 figure from testing 100+ cyclists and finding that 20-minute power runs about 5% above true 60-minute power on average. Ric Stern, who popularized the 0.75 ramp multiplier, is the first to note the real range is 72–77%, not a clean 75%.
How to Pace the 20-Minute FTP Test
The 20-minute test is accurate when paced right. It falls apart when you go out too hard.
Start at a watt target you’ve held for 10–15 minutes in training. A rider who blows up at minute 12 and limps home records a number 15–25W below their real FTP.
Step-by-step:
- Warmup: 10–20 minutes easy, then 2–3 short punches at threshold or above. You’re priming your legs, not saving them.
- First 10 minutes: controlled, slightly uncomfortable. Resist the urge to go harder.
- Minutes 10–15: if you feel strong, add 5–10W.
- Final 5 minutes: everything you have left.
- Calculate: average power × 0.95 = FTP.
ERG mode or resistance mode? Use resistance mode (or slope mode). ERG mode locks your power and fights you if you slow down — it can artificially prop up a fading rider and skew the average upward. Resistance mode keeps it honest.
A 2019 reliability study found test-retest limits of agreement of +13W to −17W with a bias of just +2W. That’s reasonable for a protocol you’ll repeat every 4–6 weeks.
The Ramp Test: Fast, Forgiving, and Biased By Your Rider Type
The ramp test is the easiest protocol to execute. Power increases by 20W every minute until you can’t hold the target anymore. Stop, record your best 1-minute average, multiply by 0.75. Done.
No pacing decisions. No blowup risk. Twenty-five minutes total.
Zwift’s standard version starts at a low base wattage and adds 20W/min. There’s a “Ramp Lite” variant that goes 10W/min, recommended if you’re under 60 kg or your FTP is below 175W.
The catch: it can read wrong depending on your rider type.
Think of W’ (W-prime) as your sprint battery — the reservoir of anaerobic energy you can burn above threshold before the lights go out. Sprinters carry 25–30+ kJ in that battery. Pure endurance riders carry 9–15 kJ.
During the ramp test, you tap that battery to push the final minutes. A sprinter squeezes deeper into oxygen debt before quitting, pushing their 1-minute peak above their aerobic ceiling — so the 75% formula overshoots their real FTP by 5–10W. Their true ratio is closer to 70% of MAP. An endurance “diesel” rider sits at the other end, often 76–78% of MAP, so the ramp test may undercount their FTP.
If your ramp test feels too high — zones feel crushing, threshold workouts fall apart at minute 30 — reduce by 5W and retest.
The 8-Minute Test: The Most Accurate Option Nobody Talks About
Two 8-minute maximal efforts. Ten minutes of easy recovery between them. Average the watts from both, multiply by 0.90.
It’s not as clean as the ramp test or as established as the 20-minute test. But in a Cycling Weekly head-to-head comparison, where one rider rode all three protocols plus an actual 60-minute TT, the 8-minute protocol came closest to true 60-minute power. It ran only 2W above the actual hour test. The 20-minute test was 5W high. The ramp test was 13W high for that rider.
The 8-minute test controls for anaerobic bias better than the ramp because you sustain a hard effort instead of sprinting to failure. It’s harder to pace than the ramp but easier than the 20-minute test.
Who should use it: experienced riders who want the most precise number, or anyone whose ramp test has felt inflated.
Which Test Should You Actually Pick?
First time testing, on Zwift or a smart trainer: Take the ramp test. It’s there, it works, and you’ll have a number in 25 minutes.
Returning after an off-season break: Research on 20 well-trained cyclists showed FTP declined only 0.67% after 8 weeks off — statistically insignificant. Your metabolic efficiency may have dropped (fat oxidation fell 21.5% in that study), so the first few weeks may feel harder than the number predicts. That’s normal.
You’ve been riding with power for 6+ months: Try the 20-minute test. You have the discipline to pace it.
Your ramp test feels inflated: Run the 8-minute protocol once to cross-check. If it comes in 8–10W lower, trust the lower number.
Warmup: It Matters Less Than You Think
There’s a study that every anxious first-timer needs to read. In 2020, Barranco-Gil and colleagues tested three warmup conditions on 15 male cyclists: a standard 45-minute pre-test protocol, 10 minutes at 60% VO2max, and no warmup at all. FTP values didn’t differ meaningfully across any of them. The correlation between FTP and respiratory compensation point held at r = 0.86–0.93 regardless.
Spin easy for 10–15 minutes. Do a couple of short punches to open the legs. That’s enough.
A Real Example: Marco’s First Test
Take a fictional but plausible rider, Marco (31, 78 kg), new to structured training. He started riding indoors in January and did his first ramp test after three weeks of base work. His best 1-minute power came in at 286W, converting to 215W FTP.
Marco’s zones felt right for the first few weeks. Then he started threshold workouts. They felt impossible — heart rate spiking by minute 15, legs gone by minute 20. He reduced his FTP by 8W based on how workouts were going.
Six weeks later he took the 20-minute test. His average: 228W. Multiplied by 0.95 = 216W FTP. Nearly identical to where he’d landed by feel. The ramp test gave him a working start, and the data confirmed it.
You Don’t Have to Test at All in Your First Month
There’s a formal alternative to every protocol above. It’s called Critical Power (CP) modeling, and it estimates your FTP from the data you’re already generating on every ride.
The math: plot your best power at several durations (2, 5, 12, and 30 minutes) and fit the curve to the formula P = W’/t + CP. That CP value is your critical power, and FTP runs about 96% of CP for most endurance riders.
AthleteOS estimates your FTP from the power data in your normal rides. No formal test required. As soon as you upload rides with power data, your training zones are calibrated. When you eventually take a formal test, AthleteOS absorbs the new number and adjusts zones instantly.
This matters most in your first four weeks. A formal test that early often ends in a poor pacing job and a low, demoralizing number. Better to train on a provisional FTP until you understand your own effort levels.
TrainerRoad’s AI FTP Detection uses the same underlying approach. It was validated on 22,000+ athletes and works entirely on sub-threshold ride data. You don’t have to go anywhere near maximum effort for the model to converge.
When to Retest
Four to six weeks into a structured training block. Not sooner. Adaptation takes time, and testing too often just adds fatigue without new information. If you’re building your aerobic base through Zone 2 work, expect the biggest gains at the 8–10 week mark, not week 3.
If workouts feel light and heart rate runs lower than usual, that’s a better signal than the calendar.
If you’re mixing lifting and cycling, read how strength and endurance training interact — concurrent loading can suppress FTP gains in the first 4–6 weeks. And if you keep landing at the top of Zone 4 on easy rides, check why you might always be in Zone 4.
Ready to set your zones without the test stress? Start with AthleteOS and let it estimate your FTP from rides you’re already doing.
One last thing: your first FTP number is not your real FTP. It’s your starting point. The number that matters is the one you record 8 weeks from now.