Beyond 20-Minute Tests: Why Real Threshold Power Requires Dynamic Fatigue Modeling

The standard prescription for cycling training has remained largely unchanged for two decades. You ride as hard as you can for 20 minutes, multiply the average power by 0.95, and call the resulting number your Functional Threshold Power (FTP). You then calculate five to seven training zones from that single anchor point and base the next twelve weeks of your training on it.

It is a highly efficient way to build a training plan. It is also fundamentally flawed for anyone past their first two years of structured training.

The 20-minute test assumes that every athlete's physiology fits a neat bell curve. It assumes that your anaerobic contribution to a 20-minute effort is exactly 5 percent. It assumes that your threshold power can be held for exactly 60 minutes. And it assumes that your power when fresh is the same power you can produce in the third hour of a race.

None of those assumptions hold true in the real world. To train accurately, you need to stop looking at threshold as a static point and start looking at the modeled power duration curve.

The flaw in the 95 percent rule

To understand why the math fails, look at two different athletes taking the same test.

Despite the identical test result, their actual thresholds are radically different. Rider A has a massive anaerobic engine. During that 20-minute test, they burned through a deep reserve of anaerobic energy, inflating the average. Their true physiological threshold—the point where lactate clearance matches lactate production—is likely closer to 265 watts. If they try to do 2x20 minute threshold intervals at 285 watts, they will fail by minute 12.

Rider B has almost no anaerobic capacity but a highly developed aerobic engine. Their 20-minute effort was entirely aerobic. Their true threshold is likely 292 watts. If they train at 285 watts, they are sitting in Zone Y—the band just below threshold that generates fatigue without delivering the specific adaptations of true threshold work.

A single test cannot separate anaerobic capacity from aerobic power. A modeled power duration curve can.

Deconstructing the power duration curve

A power duration curve looks at all your maximum efforts across every duration—from a 5-second sprint to a 3-hour steady ride—and plots them on a graph. Advanced models, specifically the WKO5 iLevels system we use, analyze this curve to isolate the different energy systems.

Instead of one flawed FTP number, the model gives you a complete physiological profile. These are the metrics that actually dictate how a session should be built.

mFTP (Modeled FTP)

This is your true threshold. It is not calculated by taking a flat percentage of a single test. It is derived by looking at the relationship between your short, sharp efforts and your long, sustained efforts. By factoring in your anaerobic capacity, the model strips away the "noise" and finds the exact wattage where your aerobic system maxes out.

TTE (Time to Exhaustion)

Threshold is an intensity, not a duration. The old definition stated that FTP is the power you can hold for one hour. In reality, TTE ranges from 30 to 75 minutes depending on the athlete and the time of year.

Knowing your TTE changes how you structure extensive threshold sessions. If your mFTP is 250 watts and your TTE is 35 minutes, a standard session of 2x20 minutes at threshold is physically impossible. You will fail the second interval. Your progression needs to start at 3x10 minutes. If your TTE is 65 minutes, 3x10 minutes is a waste of time. You need continuous blocks of 30 to 40 minutes to stimulate adaptation.

FRC (Functional Reserve Capacity)

FRC is your anaerobic battery. It is measured in kilojoules (kJ) and represents the total amount of work you can do above threshold before you completely fatigue. A track sprinter might have an FRC of 30 kJ. A steady-state time trialist might have an FRC of 12 kJ.

This dictates your intensive intervals. If we prescribe 3x15 minutes at 100–103% of threshold, an athlete with a low FRC will empty their battery quickly and struggle to finish the session. An athlete with a massive FRC might find 103% too easy because their anaerobic battery subsidizes the effort. FRC tells us exactly how high above threshold you can go, and for exactly how long.

Feeding the model: The role of max-effort probes

A model is only as good as the data it processes. If you spend three months riding exclusively in low Zone 2, your modeled curve will decay. The software will assume you have lost fitness, and your mFTP will drop artificially. To keep the model accurate, it must be actively fed.

This does not mean racing every workout. It means inserting specific, calculated probes into your training weeks to anchor the curve. If the model is missing data at the 3-minute mark, a single max-effort 3-minute interval at the end of an endurance ride provides the data point needed to calculate your FRC.

When we see Pmax capping the model, we do not rewrite the training block. We simply insert dedicated max-effort probes—three or four 10-second all-out sprints with full recovery—into a standard aerobic ride. This fixes the curve, updates the model, and ensures the threshold targets for the following week are accurate.

Durability and the 2,000 kJ mark

The final variable that a 20-minute test ignores is stamina. A standard FTP test is performed fresh, usually after a rest day and a thorough warm-up. It tells you what you can do at minute 30 of a race.

If you race half-Ironman distances, marathons, or long road races, your fresh FTP is largely irrelevant. What matters is your durability—your ability to produce power after you have already burned 2,000 or 3,000 kilojoules of work.

The power duration curve tracks your stamina metric. It reveals the decay in your power output after fatigue sets in. Two athletes might share an mFTP of 280 watts when fresh. After 2,000 kJ, Athlete A might see their threshold drop to 260 watts, while Athlete B drops to 230 watts. Athlete A will ride away from Athlete B in the final hour, despite having identical "FTP" numbers.

By modeling this decay, we can prescribe specific late-stage intervals. If your stamina is a limiter, we program threshold efforts at the end of a three-hour endurance ride, forcing the body to hold threshold power under deep metabolic fatigue.

The case for keeping the 20-minute test

With all these flaws, it is reasonable to ask why the 20-minute test still exists. The answer is simple: it is highly accessible, and for certain populations, it is sufficient.

If you are in your first year of using a power meter, a modeled power duration curve is an unnecessary complication. You do not have enough historical data to feed the model, and your fitness is changing so rapidly that finding a granular TTE is a waste of focus. For a beginner, a well-executed 20-minute field test provides a number that is close enough to establish baseline training zones.

A single test is also easier to execute mentally. Riding for 20 minutes requires focus, but it is a known quantity. Feeding a power duration curve requires a willingness to go to absolute failure across multiple different durations—30 seconds, 5 minutes, 40 minutes. It demands a higher level of athlete engagement.

However, "close enough" comes with an expiration date. Once you have built a solid aerobic base and are looking to push your physiology further, the margin of error in a 20-minute test becomes too costly. Training at 95% of a 20-minute average works until you plateau. Breaking that plateau requires knowing exactly where your threshold sits, how long you can hold it, and how much reserve capacity you have to push beyond it.