Saturday's four-hour ride felt fine on water and whatever was already in the tank — strong legs, decent split, no bonk, no drama. Dinner was normal. Sleep was normal. Sunday's hard group effort wasn't: legs that should have recovered felt hollow twenty minutes in, and the numbers came in soft everywhere it mattered. Nothing about Sunday changed. Saturday did — the ride didn't get fuelled, and the bill came due a full day later.

Section 01What the data says

A 2023 study in Medicine & Science in Sports & Exercise put this exact scenario under lab conditions. Seven trained male cyclists (VO2max ≈ 67 ml·kg⁻¹·min⁻¹) rode to exhaustion — roughly two and a half hours at 68% of VO2max, then repeated one-minute-on, one-minute-off intervals at 90% of VO2max until they couldn't hold it — once fuelled with a carbohydrate drink during the ride (1 g of carbohydrate per kg of bodyweight per hour) and once on a water-only placebo, at least a week apart, matched for total work both times (Clauss et al. 2023).

The rest of that day and the next morning were controlled too — an isoenergetic, standardised diet, identical for both conditions, so nobody could out-eat the deficit at dinner. Then, the next day, everyone rode a time trial (Clauss et al. 2023).

Worth sitting with how demanding the initial ride actually was: roughly two and a half hours at a genuinely hard aerobic pace, then repeated max-effort intervals to failure on top of it. This isn't a moderate Sunday spin — it's the kind of session a big group ride, a hard gran fondo, or the first big day of a multi-day tour or stage race actually produces. That's exactly why the finding matters: the rides most likely to be under-fuelled by accident, because "I've got the legs for it," are the same rides this study shows carry the biggest next-day bill.

The riders who fuelled with carbohydrate the day before finished that time trial 61 seconds faster — 41:49 versus 42:50 (P = 0.020) — and pedalled more efficiently doing it (Clauss et al. 2023). Nothing about the time trial itself differed between conditions. The only thing that changed was what happened to fuel on a ride more than eighteen hours earlier.

Section 02Why the comparison is hard to argue with

The design is what makes this worth taking seriously. Each cyclist rode both conditions — carbohydrate and placebo — so every rider was compared against himself, not against a different group with different genetics and training history. The two sessions were matched for total work and separated by at least a week to wash out any carryover (Clauss et al. 2023). Most importantly, the standardised, isoenergetic diet for the rest of that day and the next morning was identical in both conditions — so whatever showed up in the next-day time trial can't be explained by one group simply eating more food overall afterward (Clauss et al. 2023).

Section 03Why a fuelled ride yesterday shows up as watts tomorrow

The likely mechanism sits in blood chemistry, not just the tank. Two amino acids linked to protein breakdown — tyrosine and phenylalanine — rose through the ride and the following hours in the no-carb condition, and stayed flat when carbohydrate was on board during the ride (P = 0.018) (Clauss et al. 2023). Read plainly: without carbohydrate, the body reached harder into its own tissue to keep the exhaustive effort going, and some of that cost was still being paid the next morning.

It's worth being precise about what did and didn't move. Two other markers of whole-body protein breakdown — urinary nitrogen and 3-methylhistidine — showed no difference between conditions (Clauss et al. 2023). The honest read isn't "skipping carbs shreds your muscle." It's narrower and still useful: carbohydrate availability during a long, hard ride changes something in amino-acid metabolism that tracks with slower, less efficient riding the next day — whatever the full mechanism turns out to be.

Section 04This isn't the same problem as chronic under-fuelling

This is a different failure mode to chronic low energy availability — the slow drain where a rider hits "maintenance" calories across weeks while never actually subtracting the training cost from the number. That's a multi-week accounting problem, and it's the kind of gap that takes months of tracking to even notice. This is acute: one long or exhaustive ride, under-fuelled, and the cost shows up in a single number on a single morning — even when the day's total energy intake was otherwise identical between conditions (Clauss et al. 2023).

The two problems can obviously compound — a rider running chronically low energy availability across a training block is also more likely to under-fuel any single long ride inside it — but they're not the same finding, and they don't call for the same fix. Chronic low availability is corrected across weeks, day by day, against a rolling trend. This is corrected one ride at a time, on the day it happens, because the study's own design shows a normal recovery meal and a full night's sleep afterward weren't enough to close the gap on their own.

Section 05Stacking hard days without stacking the debt

Multi-day tours, stage races and back-to-back weekend blocks are where this finding earns its keep. If day one is genuinely long or hard and day two matters — a harder stage, the last big day of a tour, the group ride you actually want to be strong for — day one's fuelling isn't a comfort decision anymore. It's the input to day two's power number, measured directly in this trial as a next-day time-trial gap (Clauss et al. 2023). Riders who fuel conservatively on "easier" long days because they finished comfortably on water are, on this data, quietly taxing the day that actually counts.

Section 06Fuel the ride you're on, not just the day you're having

The dose in the trial wasn't exotic. One gram of carbohydrate per kilogram of bodyweight per hour works out to about 70 g an hour for a 70 kg rider — roughly two standard gels plus a strong bottle of mix, or whatever combination gets you there with what's already in your kit. It's close to what most long-ride fuelling guides already recommend for how a ride feels today. What this study adds is the reason to actually hit that number on the days you're tempted to skip it: the cost of skipping it doesn't stay in today's ride. It shows up on tomorrow's.

1. Fuel during the ride, not just after it

   The dose that worked in the trial was 1 g of carbohydrate per kg of bodyweight per hour, taken during the ride itself — not backloaded into a recovery meal.
2. Treat back-to-back hard days as one event, not two

   If Saturday's long ride and Sunday's hard effort are both on the calendar, Saturday's fuelling is an investment in Sunday's power, not just Saturday's comfort.
3. Don't expect dinner to undo it

   The placebo condition ate a normal, matched diet afterward — the next-day cost still showed up. A big recovery meal is not a substitute for fuelling the work itself.
4. Reserve this for the rides that actually exhaust you

   The effect was measured after roughly 2.5 hours of steady load plus max-effort intervals to failure — not every easy spin needs an in-ride fuelling strategy this aggressive.

Section 07Where riders misread this

Section 08Applying it with THRIVE

Section 09Bottom line

Fuelling a long or hard ride protects more than today's ride. In a controlled trial, skipping carbohydrate during an exhaustive session cost a trained cyclist real time and efficiency on a time trial nearly a full day later — after an identical, matched diet in between. If tomorrow matters as much as today, fuel today like it does.

A note on limits. This is one controlled study in seven trained male cyclists — a small, tightly-controlled sample, not a guarantee of an identical effect for every rider, every ride length, or every carbohydrate dose. The effect on next-day time-trial time was itself small in statistical terms, even though it was measurable and consistent with the efficiency and amino-acid data. None of this is licence to over-eat on easy days — it's about not shorting the fuel a genuinely long or hard ride costs you, not eating more in general.