TLDR: Carb loading for a marathon means eating a lot more carbohydrate in the 36–48 hours before a long race so your muscles start fuller than usual. The evidence that it fills the tank is strong: in recent research, more carbs meant more stored glycogen, up to about 10 g per kg of body weight a day. The evidence that a fuller tank makes you faster is weaker than most people assume. Very few studies have used a placebo, and the ones that did, with proper fuelling during the race, didn’t find a difference. My take: the mechanism makes sense, the risk is low when you do it properly, and if part of the benefit is belief, that’s still a benefit. If your race is longer than about 90 minutes, load, and do it with the right foods.
Walk past any big city marathon expo on the Saturday and you’ll find it. The pasta party. Long tables, paper plates, a mountain of penne and a garlic bread arms race. Carb loading is so baked into marathon culture that most runners do it because everyone else does it.
That’s not necessarily a bad reason, but it’s not a great one either. So let’s look at why carb loading should work, what the lab says, and the harder question of whether it actually makes you faster come race day.
What is carb loading?
Carb loading is eating much more carbohydrate than usual in the days before a long event, while training less, so your muscles store more glycogen than normal.
Glycogen is how your body stores carbohydrate. Think of it as your fuel tank. Under normal conditions you hold roughly 400 g in your muscles and another 100–120 g in your liver (Jones et al. 2026). During hard endurance exercise, glycogen is your main fuel, and those stores do eventually run down, leaving you feeling like you’ve just been hit by the blue shell on the final lap of Mario Kart. When they get low, your ability to hold pace drops with them.
Topping the tank up beyond its normal level is called glycogen supercompensation. That’s the whole point of a carb load, to fill your tanks to the brim.
A short history of carb loading
Carb loading as a practice can trace its origins to research done in Scandinavia in the 1960s, when scientists began taking muscle biopsies to see what was actually happening inside working muscle (Bergström & Hultman 1966).
The original “classic” protocol was brutal. Two exhausting exercise sessions separated by three days of very low-carbohydrate eating, followed by three days of rest and a high-carbohydrate diet. It worked: muscle glycogen ended up well above normal (Bergström et al. 1967). It also meant three miserable days of flat, carb-starved training in race week, which is not what anyone wants just before a marathon and you’ll be relieved to know it’s not something I will be advising you do either. Three days of no carbs in race week is the nutritional equivalent of tapering with hill reps.
Thankfully, later research showed the depletion phase isn’t necessary. Shorter periods of very high carbohydrate intake, without the exhausting sessions, can do the job just as well. One study had trained athletes eat 10 g of carbohydrate per kg of body weight while resting, and they substantially raised their glycogen in just 24 hours, with no extra gain from carrying on to 72 hours (Bussau et al. 2002).
The current sports nutrition consensus recommends 10–12 g/kg a day for the 36–48 hours before events longer than 90 minutes (Thomas, Erdman & Burke 2016).
How many carbs do you need to carb load?
This is where 2026 has been a big year. A research group at Liverpool John Moores University has published two papers that sharpen the answer.
The meta-analysis. Jones and colleagues pulled together 63 trials and 641 participants, all of which measured glycogen directly with muscle biopsies. Going from a moderate intake (under 6.5 g/kg a day) to a high one (over 6.5 g/kg) raised muscle glycogen substantially. In endurance-trained people, there was a clear dose-response: the more carbohydrate eaten, the more glycogen stored, with no obvious ceiling up to about 10–12 g/kg. Their practical recommendation was more than 8 g/kg a day for 36–48 hours. Loading for 72–96 hours added nothing other than perhaps the extra joy that comes alongside the sheer amount of Haribo and crumpets you consume, all in the name of performance of course.
The dose-response trial. In a separate study, 11 endurance-trained athletes who kept training, as you would in a taper, ate 6, 8 or 10 g/kg a day for 48 hours on separate occasions. 10 g/kg resulted in the most glycogen. Interestingly, 8 g/kg was no better than 6.
10 g/kg for a 70 kg runner is 700 g of carbohydrate a day. That’s a lot of food, which is exactly why the type of carbohydrate matters (more on that below).
Why your tank isn’t the same as someone else’s
The meta-analysis also showed big differences between people eating the same amount. A few factors explain some of that.
Training status. This is the biggest one. Trained people store far more glycogen than untrained people, and in one study untrained participants stored nearly twice as much after ten weeks of endurance training (Greiwe et al. 1999). It stands to reason that a 2:30 marathon runner may need more carbohydrate to fully load than a 4:00 runner of a similar size and weight, although nobody has tested this directly yet.
Body size. Recommendations in grams per kg get daunting for bigger athletes. At 90 kg, 10 g/kg is 900 g of carbohydrate a day. There is a point where appetite and gut comfort become a huge limiting factor, and there’s a sensible case for a practical absolute upper limit.
Do you need to carb load for a half marathon?
It depends on how long you’ll be running, not the distance.
- Under 60 minutes: firstly, well done for being an elite runner; secondly, you don’t need to load. Eat your normal high-carb diet.
- 60–90 minutes: “fuel up” rather than load. Around 7–8 g/kg in the 24 hours before is plenty.
- Over 90 minutes: a full carb load, 36–48 hours out is a good idea.
So a 1:25 half marathoner doesn’t need a full load, but a 2:30 half marathoner probably does. Same race, different answer. For marathons, Ironman and ultras, loading is on the table for everyone.
Does carb loading actually make you faster?
This maybe isn’t as definitive as you’d think!
A 2026 commentary from the same Liverpool group looked at the 14 studies that tested carb loading’s effect on performance. Two problems stood out.
Most didn’t look like a real race. Ten of the 14 gave no carbohydrate during exercise, and seven of those had people turn up after an overnight fast with only water. That’s not how anyone races a marathon. If you fuel well during the race, a fuller starting tank may matter less.
Almost none used a placebo. The authors classed only two as double-blind, placebo-controlled trials. Both gave carbohydrate before and during exercise, and neither found a performance benefit.
The better-known of the two is Burke and colleagues (2000). Seven well-trained cyclists were told they were testing two sports supplements for race preparation. All their meals were provided and identical across both trials. The only difference was the snacks: sports bars that added 3 g/kg of carbohydrate in one trial, and a sweetened, low-calorie drink described to them as a “carb-loading” drink in the other. The real load raised muscle glycogen. But in a 100 km time trial, with carbohydrate taken during the ride, finish times were 147.5 versus 149.1 minutes, about 1% faster, which wasn’t statistically meaningful.
That doesn’t prove carb loading is useless. The authors pointed out that 1% could matter to elite riders, and that they’d have needed around 30 participants to detect such a small effect with enough confidence. The two unblinded studies in the commentary that did fuel participants properly found benefits, but their effects were about the size you’d typically see from a placebo nutrition intervention.
The two most common carb loading mistakes
In practice, I see two mistakes far more often than anyone getting the grams slightly wrong.
1. Treating it as a free pass. Carb loading isn’t a licence to eat whatever you like. Pizza, creamy pasta, pastries and desserts feel like carb loading, but a lot of their energy comes from fat. What you end up with is a very high-calorie day, not a carb load. Your glycogen doesn’t care how full you feel; it cares how much carbohydrate you ate.
2. Trying to eat too healthily. The opposite trap. Wholegrains, beans, lentils and big piles of vegetables are great most of the time, but they’re bulky and high in fibre. Try to hit 700 g of carbohydrate from those and you’ll be painfully full, and quite possibly spending race morning in the portaloo queue. Instead try low-fibre, high-GI carbohydrate to make very high intakes more tolerable. White rice, white pasta, white bread, bagels, potatoes without the skin, cereal, juice, sports drinks and, yes, a few sweets.
So, should you carb load?
Here’s where I land.
The evidence isn’t as strong as marathon culture makes out. We know carb loading fills the tank. We don’t know for certain how much a fuller tank adds when you also fuel well during the race.
But the mechanism makes sense, and done properly, it’s a very low-risk strategy, and if some of the benefit comes from belief, from standing on the start line knowing you’ve prepared properly, that’s still a benefit. After all, if the benefit is all in your head, your head still has to run 26.2 miles.
My rule of thumb:
- If your race will take longer than about 90 minutes, carb load.
- Aim for 8–12 g/kg a day for the 36–48 hours before, while you taper.
- Get it mostly from white, low-fibre carbohydrate, and keep fat lower than usual.
- Practise it before a long training run or a lower-priority race first.
- Then fuel properly during the race. That part has the stronger evidence.
To work out your numbers and get a two-day plan, use my free Carb Loading Calculator. For race-day carbohydrate, fluid and sodium, use the Race-Day Fuel Calculator.
Common questions
When should I start carb loading for a marathon?
36–48 hours before the start. For a Sunday morning marathon, that means Friday morning or lunchtime. Starting earlier doesn’t seem to add anything.
Does carb loading make you gain weight?
Not much, if at all. In a 2026 trial, body mass at 10 g/kg a day was no different to 6 g/kg (Jones et al. 2026). If the scales do move, it’s stored fuel and water, not fat.
Do I need a depletion run before carb loading?
No. The old depletion phase isn’t necessary for trained athletes. A normal taper with a high carbohydrate intake does the job.
Is pizza good for carb loading?
In moderation, as part of a day that’s otherwise high in carbohydrate. On its own it’s too high in fat to be an efficient carb load.
Is carb loading a myth?
No. It reliably increases muscle glycogen. Whether it makes you faster when you also fuel during the race is less certain, which is why the best plan is to do both.
