Beyond the 20-Minute FTP Test: New Protocols Gain Traction

The 20-minute FTP test has dominated cycling training for nearly two decades, but a growing number of coaches and platforms are abandoning the protocol in favor of alternatives they claim are more accurate, less painful, and better suited to modern training analytics. TrainerRoad, Wahoo, and Zwift have all updated their testing protocols in recent months, while coaching software like Today's Plan and Training Peaks now offer AI-estimated FTP calculations that eliminate formal testing entirely.

The shift reflects accumulating evidence that the classic test—ride as hard as possible for 20 minutes, multiply average power by 0.95—produces inconsistent results depending on pacing strategy, motivation, fatigue, and individual physiology. A study from the University of Kent tracked 50 trained cyclists performing 20-minute tests monthly for six months and found test-to-test variability of 4-7 watts even under controlled laboratory conditions, suggesting the protocol measures performance capacity as much as physiological threshold.

The Case Against 20-Minute Testing

Traditional FTP testing assumes a linear relationship between 20-minute power and one-hour power, with the 0.95 multiplier theoretically accounting for anaerobic contribution during shorter efforts. But research shows this relationship varies significantly across individuals. Some cyclists can sustain 97-98% of their 20-minute power for an hour, while others drop to 92-93%, meaning the standard 0.95 multiplier systematically overestimates FTP for certain riders and underestimates it for others.

Pacing strategy introduces another variable. Cyclists who start too hard often blow up at 12-15 minutes, producing artificially low average power that underestimates true threshold. Conversely, conservative starters may finish with excess energy, suggesting they could have sustained higher power across the full interval. This pacing sensitivity makes 20-minute tests as much about race strategy and mental fortitude as physiological capacity.

Motivation and context matter enormously. Testing indoors on a smart trainer in your garage produces different results than testing outdoors on a favorite climb, even when controlling for environmental factors. Group rides or virtual racing on Zwift often yield 5-10 watts higher results than solo efforts, reflecting competitive drive that's hard to replicate alone.

Ramp Tests as the New Standard

Ramp protocols—where resistance increases every minute until failure—have become the default testing method on TrainerRoad and increasingly popular elsewhere. The typical structure starts at easy endurance power (50-60% of estimated FTP) and adds 20 watts per minute until the rider can't sustain target power for 30 consecutive seconds. Final one-minute average power, multiplied by 0.75, estimates FTP.

The ramp test's key advantage is removing pacing strategy from the equation. You simply hold target power each minute until you can't anymore—no decisions about when to accelerate or when to ease up, no opportunity to quit mentally before reaching physiological limits. Tests take 15-20 minutes versus 30+ for traditional protocols when including warm-up, making them less mentally daunting and easier to incorporate into regular training blocks.

Research from Stephen Seiler's lab found ramp tests produced more repeatable results than 20-minute efforts, with test-to-test variability under 3 watts for trained cyclists performing monthly assessments over four months. The protocol also appears to minimize the motivation variable—most riders naturally push until true failure during ramp tests, whereas 20-minute tests allow quitting prematurely when suffering intensifies.

The downside? Ramp tests emphasize VO2max and anaerobic capacity more than sustained threshold efforts, potentially overestimating FTP for riders with powerful neuromuscular systems but limited endurance. Sprinters and track cyclists often test artificially high on ramp protocols, then struggle to hold calculated training zones during longer intervals.

AI-Estimated FTP: No Testing Required

Several platforms now calculate FTP automatically by analyzing power data from normal training rides and workouts. Algorithms examine hundreds of efforts across different durations—30-second sprints, 5-minute climbs, 20-minute tempo efforts, 2-hour endurance rides—to model your power-duration curve. The curve's inflection point theoretically represents threshold, eliminating the need for dedicated testing.

Wahoo's SYSTM platform introduced this feature in late 2025, analyzing four weeks of training history to estimate FTP within ±3% of laboratory-tested values for 85% of users according to the company's validation study. The system updates estimates weekly as new data arrives, allowing FTP to drift upward during build phases or downward during recovery or off-season periods without requiring formal tests.

Training Peaks' AI FTP Detection, rolled out in April 2026, uses similar methodology but extends the analysis window to 6-8 weeks and weighs recent efforts more heavily. The platform flags when estimated FTP diverges from user-set values by more than 5 watts, prompting zone adjustments without forcing a test.

The obvious benefit is eliminating the need for exhausting, dreaded FTP tests that many cyclists postpone indefinitely. If the algorithm can track threshold shifts through normal training, why suffer through 20 minutes of maximal effort? The risk lies in garbage-in, garbage-out scenarios: if training volume is low or efforts are predominantly easy endurance work without threshold or VO2max intervals, the algorithm lacks data to model upper-end capacity accurately.

Duration-Curve Analysis for Individualized Zones

A more sophisticated approach treats FTP as one point on a broader power-duration curve rather than a single magic number. Platforms like Xert and WKO5 generate curves mapping maximum sustainable power across time frames from 1 second to 5+ hours. Training zones derive from multiple curve points rather than percentages of a single threshold value.

This method acknowledges that cyclists don't all scale the same way. A rider with FTP of 300 watts might hold 600 watts for 5 minutes (strong VO2max) or only 520 watts (limited top end). Similarly, they might sustain 280 watts for 90 minutes (excellent durability) or only 270 watts (threshold durability needs work). Traditional zone systems based solely on FTP percentage miss these distinctions.

Duration-curve training prescribes intervals based on time-to-exhaustion at target power rather than percentage-of-FTP. Instead of "5x5 minutes at 110% FTP," the workout becomes "5x5 minutes at your 7-minute power." This self-calibrating approach ensures intervals stress the intended physiological system regardless of how your power-duration curve is shaped.

The challenge is complexity. Most cyclists find "Zone 2" or "90% FTP" intuitive, while "maximal aerobic power at 5:30 duration" requires more explanation. Coaches adopting curve-based training report needing 2-4 weeks of education before athletes grasp the concepts and stop asking "but what's my FTP?"

What This Means for Your Riding

If you dread FTP tests and have been training on outdated zones for months (or years), ramp tests or AI-estimated values offer legitimate alternatives. Choose ramp protocols if you want simple, reproducible testing with less mental suffering. Choose AI estimation if you train consistently with adequate intensity variety and prefer avoiding formal tests altogether.

For most recreational cyclists training 6-10 hours weekly, testing every 8-12 weeks provides sufficient granularity to track fitness and adjust zones. Testing more frequently rarely adds value unless you're in a focused build phase with rapidly improving fitness. Testing less often risks training in outdated zones that compromise workout quality.

If you're working with a coach or using structured training plans, discuss testing protocols explicitly. Mismatches between how your FTP was tested and how zones are prescribed—ramp-tested FTP applied to plans designed for 20-minute values, for example—can lead to workouts that feel too easy or impossibly hard.

Ultimately, FTP is a training tool, not a measure of self-worth. Whether you test traditionally, ramp, or estimate algorithmically matters less than using consistent methodology and honestly assessing whether your zones produce productive training. If Zone 2 rides feel like tempo efforts, if threshold intervals leave you destroyed rather than fatigued, or if you're constantly adjusting intensity mid-workout, your FTP probably needs recalibrating regardless of which protocol you choose.