Chapter 23 — Sports Nutrition: Fuelling, Timing, Recovery, and the Five Supplements That Work

The Hook: Kilometre 32

Devi Raghunathan had run 32 kilometres of a marathon twice before, and both times something happened that she described in almost the same words:

"It's not that it hurts. It's that the pace just isn't there. My legs are doing the same thing and the watch says I'm slower and I can't make it stop."

She'd called it "the wall," and she'd assumed it was a mental problem, because that's what everybody had told her.

⚠️ It is an arithmetic problem, and it is one of the few things in this book with a clean, quantitative answer.

Here is the arithmetic.

The tank holds less than the race costs, and there is no way to make it bigger by very much.

⚠️ So a marathon is a problem of managing a deficit — starting full, spending slowly, and putting carbohydrate in during the race. Miss any of the three and the deficit arrives at roughly 30–32 km, because that's where the arithmetic runs out.

It is not willpower. It is a fuel gauge.

And after three chapters of adjudicating arguments, this chapter is a relief to write.

⚠️ Sports nutrition has unusually good evidence, and it's worth understanding why:

• Athletes can be measured. VO₂ max, power output, time to exhaustion, blood lactate, muscle biopsy.
• Outcomes are objective. ⚠️ A 2% improvement in a time trial is a number, not a self-report.
• The timeframes are short. A crossover trial can run in weeks, not decades.
• Participants are motivated and compliant — a problem everywhere else in nutrition.
• ⚠️ And performance is a much cleaner endpoint than "health."

Which is why Part V opens here. ⚠️ This chapter has six ✅ verdicts — more than any chapter so far, and more than most of Part IV managed between them.

🏃 Fast Track: §23.2 (the hierarchy), §23.4 and §23.6 (carbohydrate), §23.10 (the five that work), §23.15 (templates). Thirty minutes.

🔬 Deep Dive: §23.5 (train-low, honestly), §23.8 (low-carb endurance — where the evidence is clearer than the argument), §23.13 (⚠️ body composition, and the harm).

23.1 The hierarchy: what actually matters, in order

⚠️ Almost every athlete gets this inverted, and the inversion is expensive.

💡 ⚠️ Attention flows almost exactly in reverse. Rows 6 and 7 get the podcasts, the products and the arguments. Rows 1 and 2 decide the race.

Chapter 16's pattern arriving in a new setting: the cheapest, plainest, least marketable thing on the list is the one that matters most.

23.2 Energy availability: the foundation

Chapter 21 Case Study 2 introduced this and it belongs here properly.

Energy availability (EA) = (energy intake − exercise energy expenditure) ÷ fat-free mass

⚠️ It is the energy left for everything your body does that isn't the training session — hormones, bone remodelling, immune function, menstrual cycling, repair, and thought.

Reference figures commonly cited: around 45 kcal/kg FFM/day supports full function; below about 30 is where clinical consequences are consistently observed.

🔬 Claim → Evidence → Verdict

The claim: "Sustained low energy availability impairs health and performance across multiple systems."

What the evidence shows: ⚠️ This is now the organizing concept of the field. The IOC has issued successive consensus statements on Relative Energy Deficiency in Sport (RED-S) — an evolution of the earlier Female Athlete Triad — describing effects across bone, endocrine, menstrual, metabolic, immune, cardiovascular, gastrointestinal and psychological function, ⚠️ in athletes of any sex.

The evidence base includes controlled EA-restriction studies showing measurable endocrine changes within days, observational work across many sports, and a large clinical literature on the bone and menstrual consequences.

Verdict: ✅ Well supported. ⚠️ And it is the single most important thing in this chapter.

⚠️ The recognition problem, restated because it's the reason people present late:

Performance decline is one of the LAST things to go. By the time the times slow, the bone, hormones and immune system have been running a deficit for months (Chapter 21, Case Study 2).

What to watch instead: menstrual irregularity or absence (⚠️ never a normal consequence of training) · bone stress injuries · recurrent illness · persistent fatigue and poor recovery between sessions · poor sleep · low ferritin (Chapter 14) · reduced libido · frequent minor injuries · mood change.

23.3 Carbohydrate: the main lever

The dose depends entirely on training load, and this is where the ✅ evidence is.

(These ranges come from the joint ACSM / Academy of Nutrition and Dietetics / Dietitians of Canada position stand on nutrition and athletic performance, which is free and worth reading directly.)

🔬 Claim → Evidence → Verdict

The claim: "Carbohydrate availability determines performance in sustained high-intensity exercise."

What the evidence shows: ⚠️ This is among the most thoroughly established relationships in exercise physiology. Muscle biopsy studies from the late 1960s onward demonstrated that time to exhaustion at a fixed intensity tracks starting muscle glycogen. The relationship has been replicated for six decades, has a clear mechanism, and shows a dose-response.

⚠️ The reason is Chapter 6's mixing board: at high relative intensity, the proportion of energy from carbohydrate rises and fat cannot supply ATP fast enough. Fat oxidation has a ceiling in rate; carbohydrate does not, up to the limit of what's stored.

Verdict: ✅ Well supported — for sustained efforts at moderate-to-high intensity.

⚠️ The important boundary: this does NOT mean everyone needs a high-carbohydrate diet. A person training three hours a week does not need 8 g/kg. The dose is a function of the load, which is why §23.3's table is the answer and "carbs are good for athletes" isn't.

Devi's numbers, as an illustration: ⚠️ 55 kg, marathon build, roughly 90–110 km/week. 6–8 g/kg = 330–440 g/day. She had been eating around 220 g — which is well inside the normal range for a person and well outside it for her training load.

Where the carbohydrate goes, and why loading works

Worth understanding once, because it explains three separate recommendations at the same time.

Glycogen is stored in muscle and liver, and the two stores do different jobs:

Three practical consequences fall straight out of that table:

1. ⚠️ Carbohydrate loading works — several days of high intake with reduced training raises muscle glycogen above normal. The old "depletion phase" is unnecessary; modern protocols just raise intake and taper.

2. ⚠️ Breakfast before a morning session matters more than it looks, because liver glycogen is low after an overnight fast. This is a genuinely different argument from Chapter 4's breakfast verdict — that one was about weight; this is about having blood glucose for a hard session.

3. And "the wall" is specifically a muscle glycogen event, while ⚠️ the lightheaded, confused, can't-think version — sometimes called "the bonk" — is a liver glycogen and blood glucose event. They feel different, they have different causes, and in-race carbohydrate addresses both.

💡 ⚠️ This is why a marathoner can have "fresh" legs at 35 km and still slow down, and why an under-fuelled cyclist can suddenly be unable to do arithmetic. Different tank, different symptom.

23.4 Carbohydrate periodization, honestly

The most interesting open question in the chapter.

"Train low, compete high" — deliberately performing some sessions with low carbohydrate availability to amplify the training signal, while fuelling fully for hard sessions and competition.

The rationale is real: ⚠️ low glycogen availability upregulates signalling associated with mitochondrial biogenesis — AMPK, p38 MAPK, PGC-1α — which are the molecular hallmarks of endurance adaptation.

Protocols include: sleep low, train low (evening session, low-carb overnight, fasted morning session) · twice-a-day training with restricted carbohydrate between · train fasted.

🔬 Verdict: 🟡 Unclear / it depends.

⚠️ The molecular signalling response is robust. The performance outcomes are inconsistent. Some trials show benefit, some show none, and some show impaired training quality — because you cannot train as hard with low glycogen, and training quality is itself a determinant of adaptation.

The honest summary: promising, mechanistically coherent, and not established as improving competitive performance. ⚠️ It is also easy to do badly — a poorly-executed train-low programme is indistinguishable from chronic under-fuelling (§23.2), and that is the failure mode I see most often.

⚠️ If you are not being coached by someone who does this deliberately, do not do it accidentally.

23.5 During exercise: the numbers that matter

One of the clearest ✅ areas in the chapter, and one of the most under-applied.

Why the 60 g ceiling exists, and why mixing beats it:

⚠️ Glucose is absorbed via the SGLT1 transporter, which saturates at roughly 60 g/hour. Fructose uses a different transporter — GLUT5.

So a glucose-only drink hits a wall at about 60 g/hour regardless of how much more you drink. A glucose-plus-fructose mixture — commonly around 2:1 or 1:0.8 — recruits both pathways and permits oxidation rates well above that, with less gastrointestinal distress, because unabsorbed carbohydrate sitting in the gut is what causes the distress.

🔬 Verdict: ✅ Well supported that carbohydrate intake during prolonged exercise improves performance. 🟢 Probably true that multiple transportable carbohydrates permit higher oxidation rates and reduce GI symptoms.

⚠️ This is Chapter 15's ORS mechanism — co-transport exploited deliberately — and it's the second time in this book that knowing which transporter does what has produced a practical answer.

⚠️ And the part athletes skip: the gut is trainable. Tolerance for high carbohydrate intake during exercise improves with practice. Attempting 90 g/hour for the first time on race day is the single most reliable way to produce vomiting at kilometre 25. Practise it in training, at race intensity, with the exact products.

23.6 Protein: dose, distribution, and the window

Chapter 8's machinery, applied.

🔬 Claim → Evidence → Verdict

The claim: "You must consume protein within 30–60 minutes of training or the session is wasted."

Where it comes from: early studies showing enhanced muscle protein synthesis with post-exercise protein, generalized into a narrow window with a hard edge.

What the evidence shows: ⚠️ The window is hours, not minutes. Muscle remains sensitized to protein feeding for at least 24 hours after resistance exercise, and reviews examining timing against total intake have generally found that ⚠️ total daily protein and its distribution matter considerably more than proximity to the session.

Aragon and Schoenfeld's analysis of the timing literature is the standard reference for this correction and is worth reading.

Verdict: 🟠 Probably false as stated. ⚠️ 🟢 Distribution across the day is genuinely useful — roughly 0.3 g/kg across 3–5 meals beats the same total in one or two. The urgency is not.

⚠️ The exception worth knowing: if you trained fasted, or have another session in under about eight hours, then eating promptly matters — because you have a refuelling problem, not a synthesis window.

23.7 Fat, and the low-carbohydrate endurance question

⚠️ A real scientific question with an unusually clear answer, argued as though it were unresolved.

The claim: "Fat-adapted athletes can access a nearly unlimited fuel supply and outperform carbohydrate-dependent athletes in endurance events."

What's true, and it's genuinely interesting: ⚠️ Keto-adapted endurance athletes reach substantially higher rates of fat oxidation than previously thought possible. The FASTER study (Volek et al., 2016) measured peak fat oxidation rates in long-term keto-adapted ultra-endurance runners well above classical textbook maxima — a real and surprising finding.

⚠️ It was cross-sectional, comparing self-selected groups, not a randomized trial — so it establishes that the adaptation occurs, not that it helps.

And the performance question has been tested directly.

⚠️ Louise Burke's group studied elite race walkers on high-carbohydrate, periodized-carbohydrate, and low-carbohydrate-high-fat diets, with measured competition performance. The finding: LCHF increased fat oxidation as predicted — and impaired exercise economy, meaning more oxygen was needed to sustain the same speed. The high-carbohydrate and periodized groups improved race performance; the LCHF group did not.

💡 ⚠️ The mechanism is the point, and it's elegant: fat oxidation costs more oxygen per unit ATP than carbohydrate oxidation does.

At submaximal intensities that doesn't matter — there's oxygen to spare. ⚠️ At competitive intensities, where oxygen delivery is the constraint, a less oxygen-efficient fuel is a direct performance penalty.

So "fat-adapted" is real and it is a trade, not an upgrade — exactly what Chapter 21 Case Study 2 said about Devi.

🔬 Verdict: 🟠 Probably false that low-carbohydrate diets improve performance in moderate-to-high-intensity endurance events. ⚠️ 🟡 Genuinely unclear for ultra-endurance at low relative intensities, where the intensity constraint doesn't bite the same way and some athletes report doing well. And ✅ that fat adaptation itself occurs — nobody disputes that.

23.8 Hydration, in competition

Chapter 15's material, applied — and this is where the field has changed most in twenty years.

The old advice was to drink ahead of thirst on a schedule. ⚠️ That advice contributed to exercise-associated hyponatremia (Chapter 15's Priya Achterberg), which has killed athletes and which over-drinking causes.

Current thinking, in short:

1. ⚠️ Drink to thirst is adequate for most people in most events. 🟢 Thirst is a reasonably good regulator, and Chapter 15 §15.3's point stands: under-drinking has an alarm and over-drinking doesn't.

2. Losses above about 2% of body mass begin to impair performance in hot conditions — the effect is real, dose-dependent, and smaller in cool conditions than commonly claimed.

3. ⚠️ Sodium matters in long events, heat, and for salty sweaters. Sweat sodium concentration varies enormously between individuals — roughly an order of magnitude — which is why blanket recommendations fail and individual sweat testing exists.

4. ⚠️ Know your sweat rate. Weigh yourself before and after a one-hour session in typical conditions, accounting for fluid drunk. It takes one session and it replaces guessing.

🔬 Verdict: 🟢 Probably true that drink-to-thirst is adequate for most athletes in most conditions. ⚠️ ✅ Well supported that programmed over-drinking is dangerous. 🟢 that sodium replacement matters in prolonged exercise, heat, and salty sweaters (Chapter 15 §15.7).

Practical sodium, without buying anything

⚠️ Because "electrolytes" is a $4-a-sachet answer to a question that usually has a free one.

Sweat sodium concentration varies from roughly 200 to 2,000 mg per litre between individuals — about a tenfold range — and it is largely a fixed personal characteristic rather than something you train.

How to know roughly where you sit, for free:

And the practical answer for those who do need it: ⚠️ table salt. Roughly a quarter to a half teaspoon of salt, dissolved into a bottle with carbohydrate, delivers what a commercial sachet does at a fraction of the price — and Chapter 15 §15.6's co-transport point means the carbohydrate is doing useful work there too.

⚠️ Commercial products earn their premium on convenience and taste, not on composition. Which is a legitimate reason to buy them, and a different reason from the one on the packaging.

23.9 The five supplements that work

⚠️ Out of an industry worth tens of billions, this is the list. It is short, cheap, and boring — Chapter 16's pattern exactly.

Notes that matter more than the table:

⚠️ Caffeine gets 🟢 rather than ✅ here for consistency with Chapter 16 — the evidence for performance is strong, and individual response varies substantially. Habituation, sleep disruption and anxiety are real costs; and if you train in the evening, the sleep cost may exceed the performance benefit (Chapter 25).

⚠️ Nitrate's effect shrinks or disappears in highly trained endurance athletes. It's most useful for recreational and moderately-trained people — which is the inverse of who buys it.

⚠️ Beta-alanine and bicarbonate share a duration window — roughly 1 to 10 minutes — because both work on buffering capacity. Neither helps a marathon. Both can help an 800 m, a 2 km row, or a hard interval set.

⚠️ Bicarbonate causes GI distress in a substantial minority. Trial it in training, never on competition day, and split doses or use enteric forms if you're going to use it at all.

⚠️ Beta-alanine causes paraesthesia — the tingling from Chapter 16 Case Study 2 — which is harmless and is doing commercial work in every pre-workout on the shelf.

And a sixth worth naming separately: ⚠️ iron, in athletes who are deficient (Chapter 14). ✅ — and Devi is the case. This is not an ergogenic aid; it's correcting a deficiency that was destroying her performance.

⚠️ If you are drug-tested, all of the above must be third-party certified — NSF Certified for Sport, Informed Sport (Chapter 16 §16.2). Strict liability applies.

Why the effect sizes are small, and why that's the point

⚠️ Worth stating plainly, because supplement marketing and supplement evidence describe different magnitudes.

The five in §23.9 are typically worth somewhere in the region of 1–3% in the outcomes they affect — and for most of them the benefit is larger in less-trained people and smaller in elite ones.

Two readings of that, both correct:

⚠️ If you are an elite athlete, 1–3% is enormous. It is the difference between a final and a heat. This is why the five are used universally at the top of sport, and why the certification question in §23.10's contamination subsection is not paranoia.

⚠️ If you are not, 1–3% of your performance is invisible — smaller than the variation between two Tuesdays, smaller than the effect of one bad night's sleep, and vastly smaller than the effect of fixing anything in rows 1 to 4 of §23.1.

💡 ⚠️ The recreational athlete buying the elite athlete's supplement stack is buying the smallest available intervention while leaving the largest ones untouched.

And this is Chapter 16's structural point in its cleanest setting: the products with real evidence have honest, modest, well-quantified effects — which is exactly why they need marketing least and get it most.

23.10 The ones that don't

Briefly:

BCAAs — 🟠 for anyone eating adequate total protein; you are buying three amino acids you already have. · Glutamine — 🟠 · HMB — 🟡 in untrained or heavily-restricted contexts; 🟠 in trained athletes · "Testosterone boosters" — ❌ (Chapter 16) · Ketone esters — ⚗️ genuinely unresolved, expensive, and interesting · Most pre-workouts — ❌ as products (Chapter 16 §16.4; they contain caffeine, which works, at an undisclosed dose) · Collagen for tendon — 🟡 with an interesting mechanism and thin outcome data · Tart cherry — 🟡 for soreness; the performance case is weaker.

⚠️ And the general rule from Chapter 16 §16.9, which does all the work here: is the dose in the product the dose in the studies, and was it studied in people like you?

⚠️ The contamination problem, which is not optional if you're tested

Chapter 16 §16.2 established that supplements are adulterated at measurable rates. For a tested athlete this stops being an evidential question and becomes a career one.

⚠️ Anti-doping operates on STRICT LIABILITY. You are responsible for what is in your body, regardless of how it got there, regardless of intent, and regardless of whether the label disclosed it. "It was contaminated" is not a defence that returns your result.

What that implies, practically:

• ⚠️ Only use third-party certified products — NSF Certified for Sport, Informed Sport — and check the specific batch, because certification is batch-level.
• ⚠️ The highest-risk categories are exactly the ones marketed to athletes: weight loss, "pre-workout," "testosterone support," and anything promising rapid results.
• Botanical and "proprietary blend" products are the worst risk (Chapter 16 §16.4), because you cannot know what's in them and neither, sometimes, can the manufacturer.
• ⚠️ Keep the tub. If something goes wrong, the physical product and its batch number are the only evidence you will have.

⚠️ And the risk-benefit here is stark: the five supplements in §23.9 are worth perhaps 1–3% at the very top of performance. A sanction costs years. If you are tested and a product is not certified, the arithmetic is not close.

23.11 Recovery between sessions

When the next session is soon, this becomes the whole problem.

The four things, in order of how much they matter:

1. ⚠️ Refuel. If the next hard session is within about 8 hours, aim for roughly 1–1.2 g/kg of carbohydrate per hour for the first few hours. If it's tomorrow, total daily intake matters and the urgency doesn't.

2. Repair. ~0.3 g/kg protein, and then again every 3–5 hours.

3. Rehydrate. Roughly 125–150% of the fluid deficit — because you'll excrete some — with sodium, which is what allows the fluid to be retained (Chapter 15 §15.6).

4. ⚠️ Rest. Sleep is the recovery intervention with the largest effect size and the smallest market. It is not a nutrition intervention and it beats every nutrition intervention in this chapter.

⚠️ That last sentence deserves its own paragraph rather than a bullet. Sleep restriction impairs glycogen resynthesis, reduces training quality, raises perceived exertion, impairs immune function, increases injury risk, and disrupts the hormonal environment for adaptation. The effect sizes in sleep-extension studies with athletes are larger than anything in §23.9.

⚠️ And it interacts with this chapter directly: an evening caffeine dose taken for training quality can cost more in sleep than it delivered in the session — which is the single most common self-defeating supplement decision in recreational sport, and it is free to stop.

⚠️ If an athlete is doing all four badly and asks about supplements, the answer is not a supplement. This is the most common consultation in sports nutrition and the answer is almost always sleep and carbohydrate.

⚠️ Two things this chapter's evidence base is bad at

Stated up front, because §23.1 claimed the evidence here is unusually good and that claim needs its boundaries.

1. ⚠️ The literature is substantially male. A large fraction of sports nutrition research has been conducted in men, and the field has begun saying so explicitly — because menstrual cycle phase, hormonal contraception, and the physiology of the female athlete affect substrate use, thermoregulation and recovery in ways that are only now being studied properly.

What follows practically: ⚠️ the fuelling numbers in §23.3 and §23.5 are probably fine — they're scaled to body mass and training load. The confident-sounding specifics about timing, adaptation and supplement response are less transferable than they sound, and ⚠️ the energy availability material in §23.2 was substantially developed in female athletes and is the part that generalizes best in the other direction.

2. ⚠️ Almost all of it is short-term. A crossover trial runs for weeks. A career runs for decades. We know a great deal about what improves a time trial next Tuesday and much less about what a twenty-year fuelling pattern does to bone, endocrine function or longevity.

⚠️ This is the honest boundary of §23.1's claim. Sports nutrition has excellent evidence about performance over weeks. It has evidence much like the rest of nutrition — observational, confounded and thin — about athlete health over careers.

Which is exactly why §23.2 sits at the top of the hierarchy: energy availability is the one variable where the long-term health evidence is strong, and it is the one that gets sacrificed for short-term performance.

23.12 Different sports, different problems

⚠️ "Sports nutrition" is not one thing, and the biggest errors come from applying one sport's answers to another.

The recreational athlete, who is most of the readers

⚠️ A necessary corrective, because almost everything above was written about people training ten to twenty hours a week, and most people reading it are not.

If you train three to six hours a week at moderate intensity:

💡 ⚠️ The single most common error in recreational sport is importing elite fuelling into a non-elite training load — and it reliably produces weight gain, which is then blamed on the training.

Chapter 4's arithmetic doesn't care that the calories came in a gel.

23.13 ⚠️ Body composition, and where this gets dangerous

The section I'd least want skipped, and the one most sports nutrition writing handles worst.

Two things are simultaneously true and the tension between them is the whole problem:

1. In some sports, body composition genuinely affects performance. ⚠️ Pretending otherwise is dishonest and athletes know it.

2. ⚠️ The pursuit of body composition in sport causes serious, documented harm — RED-S, bone stress injuries, endocrine disruption, eating disorders — and athletes are at substantially elevated risk in weight-sensitive sports.

What the evidence supports:

• ⚠️ Lighter is not monotonically faster. The relationship is a curve with a peak, and past it performance falls — through lost power, impaired recovery, injury and RED-S. Most athletes who are "trying to get lighter" are past the peak, not short of it.
• ⚠️ Rapid weight loss impairs performance, reliably.
• Composition changes should be attempted, if at all, in the off-season, slowly, with adequate protein and resistance training, and monitored.
• ⚠️ Weight-making practices in combat sports — dehydration, sauna, restriction — carry documented serious risk, including deaths, and have been the subject of governing-body rule changes.

⚠️ The warning signs are §23.2's list, plus:

Preoccupation with food · anxiety about eating with others · training when injured or ill · using exercise to "earn" food · secrecy · and a coach or sport culture that comments on bodies.

⚠️ If you coach, note the last one: it is the modifiable one. The IOC consensus material and most governing bodies now say plainly that commenting on athletes' bodies is a risk factor, and it is the intervention available to you today at zero cost.

Chapter 34 handles this properly and I would rather a reader arrived there early than late.

23.14 Practical templates

Because the numbers above need to become food.

A hard training day — endurance athlete, ~60 kg, 2 h session

An easy day

⚠️ Carbohydrate drops to 3–5 g/kg. Protein stays. This is the part athletes get wrong in both directions — some eat the hard-day amount every day, some cut carbohydrate on easy days and arrive at the next hard session empty.

A strength athlete's day — ~80 kg, 4 lifting sessions/week

⚠️ Different problem, and worth showing because endurance templates get misapplied here constantly.

⚠️ The recurring error: lifters under-eat carbohydrate and over-focus on protein timing. Training quality falls, which is the thing that actually drives adaptation — and no amount of post-workout shake compensates for a session you couldn't complete.

Race week

23.15 Devi's race, resolved

Third appearance, and this is where her arc lands.

She ran the marathon.

⚠️ And what she reported afterwards was not that it was easier. It was that kilometre 32 didn't happen.

"I kept waiting for it. I got to 35 and thought — oh. That's just what it's supposed to feel like."

⚠️ Nothing about her training changed in the final block. She had been running that fitness for two years.

🧾 What a season of this costs

⚠️ Part V keeps Part IV's habit, because the pattern hasn't changed.

⚠️ Every item in the top block is evidenced. Every item in the bottom block is either the same compound at a markup or unsupported.

⚠️ And the biggest single saving is the one nobody markets: carbohydrate for long sessions can be made from sugar, salt and water. Glucose-plus-fructose in roughly a 2:1 ratio is, chemically, close to what table sugar becomes in your gut — sucrose is one glucose bonded to one fructose (Chapter 18 §18.2), which is why it works, and why the expensive version is expensive for reasons of taste, convenience and packaging rather than physiology.

Twelfth consecutive chapter where the best-evidenced option is the cheapest.

⚠️ How firmly I hold these

What we don't know

⚠️ Whether carbohydrate periodization improves competitive outcomes (§23.4). How high in-race carbohydrate intake can usefully go — reports of well above 90 g/hour in gut-trained athletes are emerging and the ceiling isn't settled. Whether ketone esters do anything (⚗️). And how much of sports nutrition research generalizes to women, ⚠️ because the literature is substantially male-dominated and the field has begun saying so out loud.

Spaced Review

1. (Chapter 6) Why does carbohydrate availability matter more at high intensity than low?

⚠️ Fat oxidation has a ceiling in RATE. Chapter 6's mixing board: as relative intensity rises, the proportion of energy from carbohydrate rises, because fat cannot supply ATP fast enough. ⚠️ And §23.7 adds the second half: fat oxidation also costs more oxygen per unit ATP, so where oxygen delivery is the constraint, a fat-based fuel is a direct penalty.

2. (Chapter 15) Why did programmed drinking replace thirst, and why did that get reversed?

It came from an era of concern about dehydration impairing performance and causing heat illness. ⚠️ It was reversed because over-drinking causes exercise-associated hyponatremia, which has killed athletes (Chapter 15). Chapter 15 §15.3's asymmetry is the reason: under-drinking has an alarm; over-drinking doesn't.

3. (Chapter 8) The "anabolic window" is 🟠. What's the 🟢 underneath it?

⚠️ Distribution. Roughly 0.3 g/kg across 3–5 meals beats the same daily total in one or two — because muscle protein synthesis responds to per-meal doses, not to the daily figure. The window is hours, not minutes; the distribution is real. (Exception: fasted training, or another session within ~8 hours — a refuelling problem, not a synthesis one.)

Project Checkpoint: Your Training Fuel

Component twenty-three. ⚠️ If you don't train, do Step 1 and skip to the next chapter — this checkpoint has an honest floor.

Step 1 — What's your actual training load?

Hours/week: _ · Sessions/week: _ · Hardest single session: ____ min

Step 2 — Find your carbohydrate target from §23.3, and compare.

⚠️ A gap of 100 g+ is common and is usually the whole finding.

Step 3 — Protein. Target 1.4–2.0 g/kg = ____ g/day. ⚠️ Then check DISTRIBUTION, not timing: how many eating occasions hit ~0.3 g/kg?

Step 4 — ⚠️ Estimate energy availability. (Intake − exercise expenditure) ÷ fat-free mass. Crude is fine. ⚠️ Below ~30: this is a clinician conversation, not a plan adjustment (§23.2).

Step 5 — In-session fuelling. For your longest session:

Duration _ min · §23.5 says _ g/hour · I currently take ____ g/hour

⚠️ Have I practised race-day fuelling in training, with the exact products? ____

Step 6 — Sweat rate, once. ⚠️ Weigh before and after one typical session, add fluid drunk. One session replaces years of guessing.

Step 7 — Build one race-day or long-session plan, in writing. ⚠️ Not "I'll take some gels."

⚠️ The last row is the one that determines whether the plan works. Every experienced endurance coach has the same story about an athlete who tried something new on race morning.

Step 8 — And the honest one:

⚠️ "Which row of §23.1's hierarchy am I actually working on?"

If your answer is row 6 or 7 and rows 1–4 are unaudited, that's the finding — and it's the most common finding in this entire chapter.

Next checkpoint (Chapter 24): your weight history — what you've tried, what happened, and what happened after.

Chapter Summary

⚠️ The hierarchy, and attention runs almost exactly in reverse: 1 energy availability ✅ · 2 total carbohydrate matched to load ✅ · 3 total protein ✅ · 4 fuelling during long sessions ✅ · 5 hydration and sodium 🟢 · 6 distribution and timing 🟢 · 7 supplements 🟢, perhaps 1–3%.

⚠️ §23.5's numbers: <45 min none · 45–75 min small or mouth rinse · 1–2.5 h 30–60 g/h · >2.5 h up to ~90 g/h with mixed carbohydrates. And the gut is trainable — practise it.

⚠️ §23.3's storage table explains three recommendations at once: muscle glycogen (~300–400 g) can't leave the muscle that stores it; liver glycogen (~80–100 g) maintains blood glucose and is substantially depleted overnight. Hence: loading works · breakfast before a morning session matters · and "the wall" (muscle) and "the bonk" (liver, blood glucose, confusion) are different events.

⚠️ §23.9's boundary: the five are worth roughly 1–3%. For an elite athlete that's a final versus a heat. For everyone else it is smaller than the variation between two Tuesdays — and the recreational athlete buying the elite stack is buying the smallest available intervention while leaving rows 1–4 untouched.

⚠️ And two honest limits on this chapter's evidence: the literature is substantially male, and almost all of it is short-term — excellent about a time trial next Tuesday, much thinner about a twenty-year career. Which is precisely why energy availability sits at the top: it's the one variable with strong long-term health evidence, and the one sacrificed for short-term performance.

⚠️ §23.13 is the section not to skip. Lighter is not monotonically faster — the relationship is a curve with a peak, and most athletes trying to get lighter are past it. And the modifiable risk factor available today at zero cost is: stop commenting on athletes' bodies.

The one thing to remember: ⚠️ Devi didn't get fitter. Her fitness had been there for two years. She ate enough, ate carbohydrate matched to her load, fixed her iron, and fuelled during the race — and kilometre 32 didn't happen.

What's Next

Chapter 24 takes on the hardest applied problem in the book, and the one with the largest gap between what's believed and what's established.

Why weight loss is easy and maintenance is hard. Adaptive thermogenesis, measured honestly. The Biggest Loser follow-up and what it does and doesn't show. The National Weight Control Registry — the people who kept it off, and what they actually have in common. Set points, settling points, and whether either is a useful model.

⚠️ The obesity medications, which have changed the landscape faster than any nutrition intervention in fifty years and which a book written in 2020 could not have anticipated. And bariatric surgery, which has the best long-term outcome data of anything in this chapter.