โ† Back to Blog

FTP Testing: Ramp Test vs. 20-Minute Protocol

Your Functional Threshold Power (FTP) determines every training zone in cycling. But the test you use to measure it can significantly affect the number you get โ€” and therefore the quality of your training.

What Is FTP?

FTP, as defined by Andrew Coggan (co-author of Training and Racing with a Power Meter), represents the highest power output a cyclist can sustain for approximately one hour. It corresponds roughly to the maximal lactate steady state โ€” the intensity above which lactate accumulates faster than it can be cleared.

Coggan's original protocol for estimating FTP was a 20-minute all-out effort, with FTP calculated as 95% of the average power (the 5% reduction accounts for the difference between 20-minute and 60-minute sustainable power).

The Rise of the Ramp Test

Ramp tests โ€” where power increases by a fixed amount (typically 20W) every minute until failure โ€” became popular through platforms like Zwift and TrainerRoad because they're shorter (~15 minutes), less psychologically daunting, and more repeatable. FTP is typically estimated as 75% of the best 1-minute power achieved during the ramp.

What Research Shows

Sitko et al. (2022), published in the International Journal of Exercise Science, compared ramp-derived FTP with 20-minute and 60-minute test protocols in trained cyclists. They found that the ramp test overestimated FTP by 3-7% compared to the actual 60-minute effort, while the 20-minute test was within 1-3%.

Lillo-Beviรก et al. (2022), in the Journal of Sports Sciences, showed that the discrepancy between ramp and time-trial FTP estimates was strongly influenced by an athlete's anaerobic capacity. Cyclists with high anaerobic power (sprinters, track cyclists) showed the largest overestimation from ramp tests โ€” up to 10% โ€” because the ramp protocol rewards anaerobic contribution that isn't relevant for sustained threshold efforts.

Zwift's own research team (Sanders & Heijboer, 2019) acknowledged in a paper in the International Journal of Sports Physiology and Performance that ramp-derived thresholds show higher variability than time-trial protocols and may not adequately reflect true physiological threshold in all populations.

Why It Matters

If your FTP is overestimated by 5%, every training zone shifts upward. Your "endurance" rides become threshold work. Your "threshold" intervals become VO2max efforts. Over weeks, this leads to:

  • Chronic fatigue from training too hard in supposedly "easy" zones
  • Inability to complete prescribed intervals (you assume you're having a bad day; actually your zones are wrong)
  • Reduced training quality at true high intensities because you're already fatigued

Practical Recommendations

  • For endurance-focused cyclists (road racing, gran fondos, time trials): Use the 20-minute test. The slight overestimation is well-calibrated for sustained efforts.
  • For track/crit racers with strong anaerobic systems: If using a ramp test, consider using 72-73% of peak 1-minute power instead of 75%.
  • For all cyclists: Validate your FTP by attempting a 40-60 minute sustained effort at your calculated FTP. If you can't hold it, your FTP is set too high.
  • Retest every 4-6 weeks to track progress and adjust zones accordingly.

Key Takeaway

Ramp tests are convenient but can overestimate FTP by 3-10%, especially in cyclists with strong anaerobic capacity. For accurate training zones โ€” particularly for endurance events โ€” the 20-minute test remains the more reliable protocol.

References

  • Coggan, A. & Allen, H. (2010). Training and Racing with a Power Meter. VeloPress.
  • Sitko, S. et al. (2022). Comparison of different testing protocols for determining FTP. International Journal of Exercise Science, 15(4), 1011-1020.
  • Lillo-Beviรก, J.R. et al. (2022). Validity of ramp incremental test to determine FTP. Journal of Sports Sciences, 40(14), 1612-1619.
  • Sanders, D. & Heijboer, M. (2019). Functional threshold power in cyclists. International Journal of Sports Physiology and Performance, 14(2), 265-270.
โ† All Articles