Table of contents
An event lasting more than 10 hours, whether it is a gravel ultra, a 300+ km sportive or a brevet, has little in common with a 4-hour ride that simply goes on for longer. Over 4 hours, a rider with a good FTP (the highest power they can sustain for roughly an hour) can afford to start a little too hard, eat irregularly and still finish reasonably well. Beyond 10 hours those mistakes add up, and the outcome depends less and less on what you can produce when fresh and more and more on how much of it you still have in hour eight.
This article focuses on single-day events. At the end there is a section on what changes when the riding is spread over several consecutive days, as in the Titan Desert.
How many calories you burn in 10 hours: the deficit you will have to manage
The quickest way to see why a 10-hour event is a different problem is to run the numbers. Take a 75 kg rider with an FTP of 260 W who completes the event at an average power of 145 W. The mechanical energy they produce is power multiplied by time: 145 W × 36,000 s = 5,220,000 J, or roughly 5,220 kJ.
There is a handy shortcut for converting mechanical kJ into kcal burned. A trained cyclist’s gross efficiency is usually between 20 and 25%, and one kcal equals 4.184 kJ, so the two factors almost cancel out: 1 mechanical kJ is roughly 1 metabolic kcal. Our rider will therefore burn around 5,200 kcal from pedalling alone.
At that intensity, around 55% of FTP, somewhere between 40 and 60% of the energy usually comes from carbohydrate, depending on the rider’s level and the terrain. That is 2,100 to 3,100 kcal, or 520 to 780 g of carbohydrate. By contrast, a trained, well-fuelled cyclist stores in the order of 500-600 g of glycogen across muscle and liver, and not all of it is available for pedalling: glycogen stored in the arms or trunk cannot be handed over directly to the legs. The better your aerobic base, the larger the share of fat you oxidise at a given power and the less glycogen you use, as we saw in the article on lactate.
Without eating, your carbohydrate balance would go negative several hours before the finish, and you do not need to empty the tank completely to feel it: once glycogen runs low, the power you can sustain already starts to drop. That said, this expenditure is not a fixed figure that you simply cover by eating the same amount. When carbohydrate intake during the event goes up, so does the share of energy drawn from carbohydrate, while the share coming from fat goes down. The 520-780 g calculated above therefore reflects a low-intake scenario, and if you eat more, actual carbohydrate use shifts upwards. That is why 70 g/h over 10 hours, about 700 g, is not excessive: much of what you eat is oxidised directly during the effort, and its main effect is to spare glycogen, especially liver glycogen, which is what keeps blood glucose up in the final hours. In events like these, nutrition stops being an add-on to performance and becomes one of its preconditions. I covered amounts, types of carbohydrate and how to build up tolerance to more than 90 g/h in the article on nutrition in cycling.
Durability: what matters is your FTP in hour eight
The values we normally use to describe a cyclist, such as FTP, VO2max (maximal oxygen uptake) or thresholds, are measured fresh. The problem is that they do not stay stable over a long effort. In 2021, Maunder and colleagues proposed the term durability to describe when these parameters start to deteriorate during prolonged exercise and by how much [1], and it has since become one of the most active lines of research in cycling.
Data from professionals point the same way. Van Erp and colleagues found that the most successful riders were those who best maintained their power as work accumulated [2]. In a 10-hour event, where any amateur racks up several thousand kJ, this factor matters even more.
A particularly telling study comes from Stevenson and colleagues [3]. They had trained cyclists and triathletes ride for two hours at moderate intensity (90% of the power at their first ventilatory threshold, VT1) and measured the threshold before and after. Power at VT1 fell by around 10%, and the heart rate associated with that threshold rose from about 142 to about 151 beats per minute. In practical terms: if your fresh VT1 is 200 W, after a few hours of riding it may be close to 180 W, and the 195 W that were zone 2 at the start now sit above the threshold. And that is after just two hours. By hour six or eight of a long event, it is reasonable to expect an equal or larger shift.
Durability is also quite individual. Two riders with the same FTP can behave very differently after five hours, and for this kind of event it is more useful to know that difference than to fine-tune your FTP by a few watts. A simple way to estimate it is to compare your best fresh 20-minute effort with one done after three hours of endurance riding. If you hold 280 W fresh and 252 W after three hours, you have lost 10%. Another rider with the same 280 W who holds 266 W has lost 5% and, for a 10-hour event, is better prepared even though both have the same FTP.
Another practical marker is aerobic decoupling: how much heart rate rises relative to power between the first and second half of a steady ride. Values below 5% on long rides are usually taken as a sign of a solid aerobic base, and Intervals.icu calculates it for every activity.
Pacing: many long events are lost in the first two hours
If sustainable power drops as the hours go by, the obvious question is how much each watt spent early on really costs. A study of professional cyclists by Mateo-March and colleagues gives a fairly clear answer [4]. For the same amount of accumulated kJ, work done above critical power (roughly, the boundary between intensities you can stabilise and those you cannot) impaired subsequent performance, whereas work done below it barely affected it. After 7.5 kJ/kg accumulated above critical power, critical power itself dropped by as much as 16%. In other words, not all kJ cost the same, and those spent above threshold in the early hours are paid back later at a disproportionate price.
With this in mind, I set pacing with one limit for the event as a whole and another for the climbs. For the whole event, an intensity factor (IF, normalised power divided by FTP) of 0.60-0.70, closer to 0.60 the longer the event or the less experienced the rider. Normalised power is an estimate of the real physiological cost of a ride with changes of pace, and for our rider with an FTP of 260 W that range works out at 155-180 W. On the climbs, a ceiling that should not exceed 75-80% of FTP during the first half, which means 195-210 W in this case.
The mistake I see most often is the exact opposite: riding the first climb at 90-95% of FTP because the legs are fresh, the group is pushing and it does not feel hard. You barely notice that effort at the time, but according to the data above it is one of the most expensive from hour six onwards.
Power targets should not stay fixed for the whole event either. If threshold drops by around 10% within a few hours, holding the same watts in hour seven as in hour two means riding proportionally harder. I would lower the ceilings by 5-10% from halfway onwards and, once power and heart rate start telling different stories, give more weight to perceived effort and the talk test. Heat makes this reading even harder, because it pushes heart rate up without any change in power, as I explained in the article on cycling in the heat.
| Reference | % of FTP | Rider with a 260 W FTP |
|---|---|---|
| Normalised power for the whole event | 60-70% (IF 0.60-0.70) | 155-180 W |
| Climbing ceiling, first half | 75-80% | 195-210 W |
| Climbing ceiling, second half | 5-10% below the previous one | 175-200 W |
Training: what changes compared with preparing for a four-hour ride
The foundation is the same: a strong aerobic engine, built mainly through volume at low and moderate intensity. What changes is the weight each session carries within the week, and the appearance of capacities that barely matter over 4 hours.
The long ride becomes the most important session of the week. The usual approach is to build from 3-4 hours up to 6-7 hours over 8 to 12 weeks, without needing to reach 10 hours in training. A 10-hour ride can cost two or three days of recovery, and that cost rarely makes up for what it adds over a 7-hour ride. What does make sense is a 7-8 hour dress rehearsal around 3-5 weeks before the event, with the same equipment, the same food and the planned pacing. Another strategy that works very well is to split that volume over two consecutive days, for example 5 + 5 or 6 + 6 hours. On the second day you start with glycogen only partly restored and with residual fatigue in your legs, neck and contact points, so much of the riding feels more like the last hours of the event than a fresh endurance ride, and recovery is usually easier than after a single 10-12 hour ride. What it does not fully reproduce is continuity: eating in hour nine or holding your position without getting off the bike can only be rehearsed on a genuinely long ride. How to fit that progression into your week depends heavily on how many hours you have, and it is exactly what I adjust with each rider in my training plans.
When it comes to training durability, the evidence is still limited and there is no consensus protocol. What makes the most physiological sense, and what I tend to use, is placing quality work on top of prior fatigue: for example, 3 × 15 minutes at 92-98% of FTP after three hours of endurance riding, rather than at the start of the ride. Another option is to split the intervals: 2 × 10 minutes at the same percentage at the start of the ride, several hours of endurance riding, then the same 2 × 10 minutes at the end. Beyond the training stimulus, this version gives you a direct measure of your durability, because comparing power, heart rate and perceived effort between the two blocks shows how much you lose over the hours and whether that loss shrinks from one month to the next. In both cases, the main goal is not to raise your FTP, but to make sure your FTP holds up when you already have hours in your legs.
Then there is the gut, which can also be trained. Cox and colleagues found that after 28 days of training with high carbohydrate availability, the amount of ingested glucose oxidised during 100 minutes of cycling rose from about 55 to about 64 g [5], and Jeukendrup’s review applies the same principle to gut tolerance [6]. In practice, that means using your long rides to eat what you will eat on race day, in the amounts you will eat it, rather than discovering mid-event that from hour six one more gel upsets your stomach.
What the power meter does not measure: position, contact points and decisions
In events like these, a large share of DNFs has nothing to do with metabolism. A neck that can no longer hold the position, numb hands, a saddle that becomes unbearable after six hours or a lower back that starts to complain end more rides than you might think. A position that works perfectly for 4 hours will not necessarily work for 10, and the only reliable way to find out is to test it on your long training rides.
On gravel, vibration adds to all this. As I explained when analysing Paris-Roubaix, absorbing impacts for hours causes fatigue in the arms, shoulders and trunk that does not show up on the power meter, but does show up in your ability to hold your position and control the bike. Tyre pressure and width have a direct effect on how much punishment you accumulate.
Finally, decisions. With low blood glucose and many hours in the saddle, mood and the ability to make simple decisions both deteriorate, and one of the first things to slip is eating. That is why I would not leave nutrition to how you feel. A reminder on your bike computer every 20-30 minutes is not an elegant solution, but it works.
The final two weeks: arriving fresh and fully fuelled
Reducing training load before the event (tapering) is one of the best-supported strategies. The meta-analysis by Bosquet and colleagues found that the most effective approach was a period of about two weeks in which volume is cut by 41-60% while intensity and session frequency are maintained [7]. For a 10-hour event, that means your last truly long ride should come about two weeks out, and that in the final days short sessions with a few blocks at race pace make more sense than long rides to check that you are in shape.
In the two or three days before the event, the focus shifts to carbohydrate loading. The joint position statement of the ACSM (American College of Sports Medicine) recommends 10-12 g of carbohydrate per kg of body weight per day for the 36-48 hours before events lasting more than 90 minutes [8]. For a 75 kg rider, that is 750-900 g a day, an amount that is hard to reach by eating as usual. In practice it usually means cutting back on fibre and fat and relying on carbohydrate-dense foods such as rice, pasta, white bread or carbohydrate drinks, something worth trying out before race week.
If it is several days in a row: recovery becomes part of the equation
In a stage race such as the Titan Desert, each day is usually shorter than a single-day ultra, often 5 to 8 hours for an amateur, but the nature of the problem changes. Reaching the finish in one piece is no longer enough: you have to arrive in a state to start again the next morning. I analysed a real case of preparing for the Titan Desert in this article.
That affects pacing first. Day one is when it is easiest to overdo it, with fresh legs and start-line adrenaline, and it is also the day whose cost you carry for the rest of the week. I would ride it even more conservatively than a single-day event of similar length.
Above all, it affects recovery between stages. Restoring glycogen from one day to the next requires eating considerably more than your appetite asks for, and appetite is often blunted by heat and fatigue. For days with more than 4-5 hours of exercise, the ACSM puts daily needs at 8-12 g of carbohydrate per kg, and when the time until the next effort is short it recommends 1-1.2 g/kg per hour for the first four hours [8]. For a 75 kg rider, that is 75-90 g per hour for the four hours after crossing the line, something that, with a blunted appetite, usually only happens if it is planned in advance.
References
[1] Maunder E, Seiler S, Mildenhall MJ, Kilding AE, Plews DJ. The importance of ‘durability’ in the physiological profiling of endurance athletes. Sports Med. 2021;51(8):1619-1628.
[2] van Erp T, Sanders D, Lamberts RP. Maintaining power output with accumulating levels of work done is a key determinant for success in professional cycling. Med Sci Sports Exerc. 2021;53(9):1903-1910.
[3] Stevenson JD, Kilding AE, Plews DJ, Maunder E. Prolonged cycling reduces power output at the moderate-to-heavy intensity transition. Eur J Appl Physiol. 2022;122(12):2673-2682.
[4] Mateo-March M, Leo P, Muriel X, Javaloyes A, Mujika I, Barranco-Gil D, et al. Is all work the same? Performance after accumulated work of differing intensities in male professional cyclists. J Sci Med Sport. 2024;27(6):430-434.
[5] Cox GR, Clark SA, Cox AJ, Halson SL, Hargreaves M, Hawley JA, et al. Daily training with high carbohydrate availability increases exogenous carbohydrate oxidation during endurance cycling. J Appl Physiol. 2010;109(1):126-134.
[6] Jeukendrup AE. Training the gut for athletes. Sports Med. 2017;47(Suppl 1):101-110.
[7] Bosquet L, Montpetit J, Arvisais D, Mujika I. Effects of tapering on performance: a meta-analysis. Med Sci Sports Exerc. 2007;39(8):1358-1365.
[8] Thomas DT, Erdman KA, Burke LM. American College of Sports Medicine Joint Position Statement. Nutrition and athletic performance. Med Sci Sports Exerc. 2016;48(3):543-568.
