The Fitness-Fatigue Model: Performance as Fitness Minus Fatigue
Quick Intro
The fitness-fatigue model proposes that each training session builds two things at once: fitness, which raises performance and fades slowly, and fatigue, which lowers performance and fades quickly. Performance on any day is the difference between the two. The model was introduced by Eric Banister and colleagues in the 1970s and is also known as the impulse-response model.
The Idea in Plain Terms
Imagine a hard session. It leaves you more fit, but also tired. Right afterward the fatigue is larger than the fitness gain, so you perform worse. As days pass, fatigue fades faster than fitness, and performance rises above the starting level before fitness itself slowly declines.
The model captures this with a simple relationship:
| Quantity | Meaning |
|---|---|
| Fitness | A positive effect of training that decays slowly, over weeks |
| Fatigue | A negative effect of training that decays quickly, over days |
| Performance | Starting level plus fitness minus fatigue |
Each training session adds a “dose” to both. Because the fatigue term is typically larger per unit but fades faster, a series of sessions produces a pattern: heavy training lowers performance while fitness accumulates, and reducing training lets fatigue fall while fitness remains.
This offers a mathematical way to understand why a taper works. By cutting training in the final days, fatigue drops and fitness holds, so performance rises on the day. It also explains why a single hard session on the day before an event hurts.
The idea is related to supercompensation, but differs in a key respect. Supercompensation describes one session’s recovery curve, while the fitness-fatigue model sums the effects of many sessions over time.
A Worked Example
A runner builds up training in the following way, as an illustration:
| Period | Training | Fitness | Fatigue | Performance |
|---|---|---|---|---|
| Weeks 1 to 4 | Hard build | Rising | High | Flat or slightly down |
| Weeks 5 to 6 | Reduced load | Slightly down | Falling fast | Up |
| Race day | Rested | Near peak | Low | Highest of the season |
The same pattern is what a coach tries to produce with a taper.
What the Research Shows
The model appears in Banister’s 1975 work and a 1976 paper by Calvert and colleagues applied to swimmers. In those studies the model fit training and performance records for individual athletes reasonably well. It has since been used to estimate when an athlete will peak, and in commercial training software.
The limits are known. A 2006 review by Hellard and colleagues assessed the limitations of the Banister model and found problems: the parameters vary between athletes and over time, the fits are unstable with few data points, and the quantities called fitness and fatigue are not directly measured in the body. Subsequent versions have added more terms to address these concerns.
A 2025 statistical analysis in Scientific Reports reported that the fitness and fatigue parameters could not be reliably separated and that the fatigue term did not improve predictions, which adds to the doubts about fitting the model to one person. That is a single recent paper. So the model is a useful way to think about training, with imperfect accuracy as a prediction tool.
Who It Suits
This framework helps:
- Coaches planning tapers and peaking.
- Athletes who use training load software.
- Students of periodization and recovery.
How to Use It
- Think of every session as adding fitness and fatigue.
- Expect to feel flat while fitness builds during hard blocks.
- Plan rest before key events so fatigue clears.
- Avoid a large cut in volume too early, since fitness will fade.
- Treat software predictions as estimates.
- Combine the model with how you feel.
Common Misunderstandings
It is not a measurement of actual physiology.
It is not precise enough to pick an exact peak day.
It is not the same as supercompensation.
FAQ
Why do I feel worse during hard training? Fatigue is temporarily larger than fitness gains.
How long does fitness last? Over weeks, which is much longer than fatigue.
Is it used in practice? Yes, in elite sport and in training platforms.
Can I calculate it myself? Software can, and it needs several weeks of data.
Where This Comes From
Banister EW, Calvert TW, Savage MV and Bach T, 1975, a systems model of training for athletic performance (Australian Journal of Sports Medicine, 7, 57 to 61). Calvert TW and colleagues, 1976, a systems model of the effects of training on physical performance (IEEE Transactions on Systems, Man, and Cybernetics, 6, 94 to 102). Hellard P and colleagues, 2006, assessing the limitations of the Banister model in monitoring training (Journal of Sports Sciences, 24, 509 to 520).
Related to: supercompensation, tapering, deload weeks and the general adaptation syndrome.