Abs Are Made in the Kitchen.
Health Isn't.

Evidence grades used here: strong trial evidence, mixed, mechanistic and untested, or none.

My starting point

I've trained in boxing and athletics since I was 18 and I still train. So I'm not neutral on the exercise side, and you should read the exercise sections knowing that.

Two questions, not one

"Abs are made in the kitchen" and "you can't out-train a bad diet" are both answers to one question: how do I lose fat? Most of the people I see actually need an answer to a different one: how do I get healthier, and stay that way? Once you treat fat loss and health as two outcomes, the research stops contradicting itself.

The kitchen camp is right about fat loss

Grade: strong trial evidence that exercise alone underdelivers on weight; mixed on why.

When you exercise more, your body claws some of it back. In the largest dataset of measured energy expenditure in adults (1,754 people), only about 72% of the extra calories burned through activity showed up as extra calories burned that day. The other 28% was compensated for, and people carrying more fat compensated more (Careau et al., 2021).

The best supervised trial shows what that looks like in practice. In E-MECHANIC, 198 adults were randomised to supervised exercise for 24 weeks, with energy expenditure measured properly (doubly labelled water). They lost less than predicted, and 42% of the lower-dose group and 24% of the higher-dose group lost no weight or gained (Martin et al., 2019).

Why it happens is still argued about. In that trial, people ate more. In a sub-study of the same trial (42 people), energy used while sitting quietly in a metabolic chamber fell by about 4%, with no change in eating (Broskey et al., 2021). In a third, neither changed (Flack et al., 2018). What replicates is the shortfall. Why it happens doesn't. Anyone who tells you confidently that it's "because you eat more" or "because your metabolism slows" is going beyond the data.

And the average hides the person. Across 61 studies, average compensation was 18%, but the spread was enormous, plus or minus 93% (Riou et al., 2015). In one 12-week supervised programme, people doing the same training changed weight by anything from −14.7 kg to +1.7 kg (King et al., 2007). The average tells you almost nothing about you.

The exercise camp is right about health, and more than they know

Grade: strong observational evidence (large cohorts); not randomised.

Pooling 20 studies and nearly 400,000 observations, people who were obese but fit did not have a significantly higher risk of dying than normal-weight fit people. Normal-weight but unfit people had nearly double the risk (Weeldreyer et al., 2024). Across 37 cohorts and more than 2.2 million people, every one-step increase in fitness (1 MET) went with about 11% lower risk of death (Laukkanen et al., 2022).

The most encouraging number I know: muscle-strengthening activity, on its own, was linked to 10–17% lower risk of death, heart disease, cancer and diabetes, with the biggest benefit at 30–60 minutes a week (Momma et al., 2022). Thirty to sixty minutes. A week.

These are observational studies. Fitter people differ in other ways, and some of the link runs backwards (people who are already ill are less fit). But the pattern is large, consistent and has held up across very different populations.

The finding that settles the argument

Fitness protects you against dying early. It doesn't protect you against metabolic disease in the same way. A review of 36 studies found that people with a high BMI who were fit had lower risk of death than people of normal weight who were unfit, but a high BMI with high activity still carried more risk of type 2 diabetes and of diabetes and heart risk factors than a normal BMI with low activity (Fogelholm, 2010).

That's why both camps can quote real evidence and still talk past each other.

Two caveats travel with it. Fogelholm says the conclusions may not hold above a BMI of 35. And in UK Biobank (77,169 people), when reverse causation was handled more strictly, the fit-but-heavier advantage held in men but not in women (Tarp et al., 2021). "Fat but fit" is real, but it isn't a free pass.

Where the kitchen camp goes wrong: unsupervised restriction

Grade: modelling and biochemical studies, consistent; mixed overall.

Cutting calories cuts nutrients unless the diet is planned. Modelling US national survey data for adults with overweight or obesity, a 30% calorie cut left 75–91% of people below the average requirement for vitamin A, calcium and magnesium, and 25–40% below it for protein, vitamin B6 and zinc (Fulgoni et al., 2024). That's a model, not measured blood levels, and four of its six authors work for Abbott Nutrition, which sells nutritional products for people on restricted diets. A small study that did measure (98 adults in Glasgow) found people following low-carbohydrate diets had markedly lower red-cell thiamine (vitamin B1) than controls, and supplements didn't fully close the gap (Churuangsuk et al., 2024).

And the fair side of it: in the CALERIE trial, calorie-restricted diets planned and taught by dietitians, in adults without obesity, were as nutritious as, or better than, people's own diets beforehand (Racette et al., 2022). So restriction isn't the problem. Unplanned restriction is. That's a stronger and more honest claim than "diets starve you of nutrients".

Usain Bolt and the chicken nuggets

The story is true. In Beijing in 2008, Usain Bolt has said he ate chicken nuggets almost exclusively, reportedly around 100 a day for ten days. He won the 100 m in 9.69 and the 200 m in 19.30, both world records.

It's true and it tells you nothing about your diet, for three reasons:

1. Survivorship. We know the story because he won. Nobody writes up the athletes who ate badly and ran badly.
2. One person, ten days, a 21-year-old elite sprinter. Performance on a given week and health over ten years are different outcomes, the same distinction this whole piece turns on.
3. His reason wasn't nutrition. He has said local food upset his stomach, so he ate the one thing he trusted. It's a story about avoiding risk before a race, retold for years as a story about diet.

What actually works, for both outcomes

The two camps are right about different outcomes, and together they beat either one.

• Lift something, twice a week. It's the best-value health habit in this piece, and in exercise trials it appears to limit how much of the effort your body claws back (Pontzer et al., 2026). Grade: the health benefit is strong observational evidence; the compensation effect is a recent review's reading of the trials, so mixed.
• Eat enough protein, set by your body weight rather than by a ratio.
• Keep any calorie deficit modest, and plan it so the nutrients don't fall away with the calories. Pairing hard cardio with heavy restriction is the combination the body resists most, according to the same review.
• Choose the plan you'll keep. In a year-long trial of four popular diets, weight loss tracked how well people stuck to the diet (r = 0.60), not which diet it was (r = 0.07) (Dansinger et al., 2005).

The line I'd want you to take away: the outcome you want most, fat loss, is the one that moves slowest and least predictably. The outcomes you're ignoring move faster, are better evidenced, and protect you in ways the scale can't show.

Where testing fits

Testing won't tell you which camp to join. It can tell you whether a restricted diet has already cost you something, and whether there's a reason weight isn't moving that has nothing to do with willpower:

• Blood chemistry: ferritin, B12, folate, vitamin D, thyroid (TSH and free T4), fasting glucose and insulin. Blood testing options
• A food record for two weeks before you change anything, so you know what you're actually eating.
• A measure that isn't the scale: waist, a strength test, how far you can walk in six minutes.

Declared interest: testing is what I sell. No trial has shown that testing before dieting improves weight loss. What it does is stop you dieting your way into a deficiency nobody checked for.

References

1. Broskey NT, et al. Effect of aerobic exercise-induced weight loss on the components of daily energy expenditure. Med Sci Sports Exerc 2021;53(10):2164–2172. PMID 34519717. doi:10.1249/MSS.0000000000002689
2. Careau V, et al. Energy compensation and adiposity in humans. Curr Biol 2021. PMID 34453886. doi:10.1016/j.cub.2021.08.016
3. Churuangsuk C, et al. Low thiamine status in adults following low-carbohydrate / ketogenic diets: a cross-sectional comparative study. Eur J Nutr 2024;63(7):2667–2679. PMID 38967675. doi:10.1007/s00394-024-03459-y
4. Dansinger ML, et al. Comparison of the Atkins, Ornish, Weight Watchers, and Zone diets for weight loss and heart disease risk reduction. JAMA 2005;293(1):43–53. PMID 15632335. doi:10.1001/jama.293.1.43
5. Flack KD, et al. Energy compensation in response to aerobic exercise training in overweight adults. Am J Physiol Regul Integr Comp Physiol 2018;315(4):R619–R626. PMID 29897822. doi:10.1152/ajpregu.00071.2018
6. Fogelholm M. Physical activity, fitness and fatness: relations to mortality, morbidity and disease risk factors. A systematic review. Obes Rev 2010. PMID 19744231. doi:10.1111/j.1467-789x.2009.00653.x
7. Fulgoni VL, et al. Impact of simulated caloric reduction on nutrient adequacy among U.S. adults with overweight or obesity (NHANES 2015–2018). J Nutr 2024;154(9):2732–2742. PMID 39025331. doi:10.1016/j.tjnut.2024.07.022 (four authors, Abbott Nutrition)
8. King NA, et al. Individual variability following 12 weeks of supervised exercise. Int J Obes 2007. PMID 17848941. doi:10.1038/sj.ijo.0803712
9. Laukkanen JA, et al. Cardiorespiratory fitness and risk of all-cause mortality. Mayo Clin Proc 2022. PMID 35562197. doi:10.1016/j.mayocp.2022.02.029
10. Martin CK, et al. Effect of different doses of supervised exercise on food intake, metabolism, and non-exercise physical activity: the E-MECHANIC randomized controlled trial. Am J Clin Nutr 2019. PMID 31172175. doi:10.1093/ajcn/nqz054
11. Momma H, Kawakami R, Honda T, Sawada SS. Muscle-strengthening activities are associated with lower risk and mortality in major non-communicable diseases. Br J Sports Med 2022;56(13):755–763. PMID 35228201. doi:10.1136/bjsports-2021-105061
12. Pontzer H, et al. The evidence for constrained total energy expenditure in humans and other animals. Curr Biol 2026;36(4):1013–1025. PMID 41653928. doi:10.1016/j.cub.2026.01.025
13. Racette SB, et al. Nutritional quality of calorie restricted diets in the CALERIE 1 trial. Exp Gerontol 2022;165:111840. PMID 35643360. doi:10.1016/j.exger.2022.111840
14. Riou ME, et al. Predictors of energy compensation during exercise interventions: a systematic review. Nutrients 2015;7(5):3677–3704. PMID 25988763. doi:10.3390/nu7053677
15. Tarp J, et al. Fitness, fatness, and mortality in men and women from the UK Biobank. J Am Heart Assoc 2021;10(6):e019605. PMID 33715383. doi:10.1161/jaha.120.019605
16. Weeldreyer NR, et al. Cardiorespiratory fitness, body mass index and mortality: a systematic review and meta-analysis. Br J Sports Med 2024. PMID 39537313. doi:10.1136/bjsports-2024-108748