The Fuel Tank: Why Modelling Glycogen Beats FTP for Training Load
- Stéphanie Quadranti
- Mar 11
- 5 min read
Updated: 5 days ago
If you’ve had enough of hearing or talking about FTP – your own or your friends’ – we get it. It's still a go-to metric for many, but we're going to take a look and why and where it falls short.
We can agree that it works reasonably well for what it was designed to do. Functional Threshold Power provides a rough estimate of the highest power a rider can sustain for around an hour, which makes it a useful starting point for structuring workouts and comparing progress across a season. When FTP rises, riders generally become capable of sustaining higher efforts for longer.
What FTP doesn’t capture so well is how a race or long event actually plays out, or when and why you’ll reach your limit.
Imagine you’re halfway through a long road race or sportive and the pace on the climb suddenly picks up. Your effort surges well above threshold for a few minutes, the pace settles briefly, and then you’re hit with another acceleration as the gradient steepens.
You’ll know from training that each of those efforts feels manageable in isolation, and your FTP might even suggest that you have the power to follow them comfortably. And yet, after enough of these surges, you don’t feel quite so good anymore. Holding the same power becomes progressively harder, even though the numbers themselves don’t appear dramatically different.
In many cases, you haven’t reached a limit of muscular strength or cardiovascular capacity – you’ve just run out of energy.
In this article
What actually limits endurance performance in cycling?
Inside the muscles and liver the body stores carbohydrate in the form of glycogen. These glycogen reserves act as the primary fuel source for high-intensity exercise, supplying the energy required to sustain hard efforts above moderate endurance pace.
At lower intensities the body can rely more heavily on fat metabolism, which provides a large but relatively slow source of energy. As intensity rises, however, the balance shifts rapidly toward carbohydrate use. Efforts close to threshold and above draw heavily on glycogen stores, because carbohydrates can be metabolised quickly enough to support the power outputs required in racing.
The difficulty is that glycogen stores are limited. Even well-fuelled athletes carry only a finite amount of carbohydrate in the muscles and liver, and once those reserves begin to fall the ability to sustain high power gradually declines.
For endurance events lasting several hours, performance therefore depends not only on how much power a rider can produce, but also on how quickly the available energy is consumed.
Why can two riders with the same FTP perform so differently?
The limits of an FTP-centred model become clearer when comparing riders with similar threshold numbers.
Two cyclists might both test at an FTP of 300 watts and appear evenly matched in controlled conditions. In a race, however, their performances can differ significantly. One rider might remain comfortable deep into the event, responding to attacks and sustaining repeated surges, while the other begins to fade despite having demonstrated the same threshold power in testing.
Several factors can explain that difference. Riders vary in how efficiently they oxidise fat at sub-threshold intensities, which influences how quickly glycogen is consumed during long efforts. They also differ in how often they produce high-intensity spikes, and in how effectively they replenish carbohydrates during prolonged exercise.
The consequence is that two riders with identical FTP values can arrive at the same point in a race with very different levels of available energy.
What does it mean to think about endurance as a ‘fuel tank’?
One way to visualise this dynamic is to imagine the body’s carbohydrate reserves as a fuel tank.
Every hard effort draws from that tank, and the harder the effort the faster the fuel is consumed. Lower-intensity riding relies more heavily on fat metabolism and therefore drains the tank more slowly, while repeated efforts above threshold accelerate depletion.
Nutrition strategies can partially replenish carbohydrates during long rides, but they rarely replace fuel at the same rate that it is consumed during high-intensity racing. Over time, the balance between fuel use and replenishment determines how much energy remains available.
From this, we can look at endurance performance from a very different perspective. It’s as much about peak power as it is how large the tank is and how quickly it empties under different conditions. In a battle between FTP vs glycogen modelling, we'd pick our fuel tank method every time.
So, what does FTP vs glycogen modelling mean? And what changes when you start modelling glycogen instead?
Modern wearable devices and power meters have made it possible to analyse training and racing data at a much deeper level than was previously available. Instead of relying only on periodic threshold tests, training systems can now examine how riders distribute effort across intensities and how those patterns influence fatigue over time.
When carbohydrate availability is incorporated into that analysis, training can be framed around the dynamics of energy use rather than a single static number (FTP). Modelling glycogen depletion allows coaches and coaching systems to estimate how quickly fuel reserves are consumed during different types of efforts and how training sessions influence that process.
This approach helps riders understand not only how much power they can produce, but also how long they can sustain the demands of a particular event.
Can you actually train your fuel tank?
You’re probably already familiar with training zones. Zone 2 in particular has become something of a fixture in endurance training discussions, often described as the foundation for building aerobic capacity and improving fat metabolism.
The basic idea is well understood. Riding at lower intensities encourages the body to rely more heavily on fat as a fuel source, which helps preserve glycogen for harder efforts later in a ride or race. Over time, this adaptation allows riders to sustain long efforts more efficiently and delay the point at which carbohydrate reserves become limiting.
What’s less obvious is how much of a particular zone a rider actually needs, or when it should appear within a training block. Standard plans typically prescribe a fixed number of endurance rides, tempo sessions or interval workouts each week, regardless of how an individual rider is responding to those sessions or how their energy use is evolving over time.
In reality, the optimal balance between endurance work, threshold training and high-intensity efforts varies considerably between riders. Some athletes naturally rely more heavily on carbohydrates and therefore benefit from a greater emphasis on endurance work that improves fat oxidation. Others may already have a well-developed aerobic base and instead need more work at higher intensities to improve their ability to repeat hard efforts without rapidly depleting glycogen.
So, how can I train my personal fuel tank?
In practice, this is where adaptive training platforms become useful. Inside Topp, the fuel tank model uses training files and wearable data to estimate how different power levels draw on your available energy reserves. Instead of prescribing a fixed distribution of training zones, the system adjusts sessions based on how efficiently you’re using that fuel and how fatigue is building over time.
Training then becomes less about following a standard schedule of zones and more about developing the metabolic adaptations that allow you to efficiently manage energy across long races and events.
If you’re curious to explore that in your own training, you can create a free Topp account and see how the fuel tank model works with your data!
Comments