Diet or Exercise for Weight Loss? The Science-Backed Answer

Disclosure: This is published by Slim Fit Nutrition, a company that sells meal plans, so read it with that in mind. Every number below is attached to a paper you can look up, and where the evidence runs against the thing I sell, it says so.

Diet or exercise, and where the 80/20 rule came from

You have almost certainly been told that weight loss is 80% diet and 20% exercise. It is on gym whiteboards and in most of the advice aimed at women who want to lose weight. I used to repeat it here myself.

I went looking for the study behind it. There is not one. No trial and no meta-analysis splits weight loss into a ratio like that, and the question does not have that shape: you cannot assign a percentage of one result to two things a person does at the same time. The paper this article used to cite for it, an editorial in the British Journal of Sports Medicine in 2015, is an opinion piece rather than a study, and it never contained an 80/20 figure at all.^[9]

Here is what the research does say, which makes a better article anyway. Johns and colleagues, writing in the Journal of the Academy of Nutrition and Dietetics in 2014, pooled the randomised trials that compared diet alone against diet plus exercise inside the same studies.^[1] At three to six months the two were level, 0.62 kg apart, nowhere near significant. At twelve months diet plus exercise was ahead by 1.72 kg. Against exercise-only programmes, the combined approach was more than 5 kg ahead at three to six months.

That is the whole argument in one paragraph. Early on, adding training does not make the scale move faster. Later on, it does.

The early part is arithmetic. Food changes are cheap in time and easy to repeat; the exercise that would match them is neither.

Those are estimates, not measurements, and the burn figures shift with body size. The training rows are gross burn, so knock off the 50 or so calories you would have burned sitting still through that run before you set it against a food row. The shape still holds: the food rows cost a decision, the training rows cost an hour you may not have.

What exercise actually buys you

Body composition rather than scale weight

Bellicha and colleagues, in Obesity Reviews in 2021, brought together 12 systematic reviews and 149 studies of exercise without a diet change.^[2] Exercise without a diet change produced roughly 1.5 to 3.5 kg of weight loss. Modest, and it is why the scale disappoints women who train hard and change nothing else. What the same programmes did produce was a shift in what the body is made of, including the visceral fat you cannot see.

A small, real bump in resting metabolism

MacKenzie-Shalders and colleagues, in the Journal of Sports Sciences in 2020, meta-analysed 18 trials measuring resting metabolic rate.^[3] Aerobic training on its own did not raise it to any significant degree. Resistance training did, by about 96 kcal a day, with a confidence interval running from 45 to 147. Ninety-six calories is a slice of bread. It is worth having, it is not a rescued metabolism, and I will come back to why that distinction matters.

The things food cannot do

Nothing you eat improves your cardiorespiratory fitness, loads your bones, or gets you up two flights of stairs without stopping. Those come from training or they do not come at all, which is a fair reason to keep training through the months when it is not moving the scale.

The metabolism story, told straight

You will read that exercise counteracts metabolic slowdown. The trials do not support it, and it is the kind of claim that sets a woman up to blame herself when it does not happen.

The slowdown itself is real. Fothergill and colleagues, in Obesity in 2016, followed contestants from a televised weight loss competition and measured resting metabolic rates around 500 kcal a day below what their body composition predicted, still there six years later.^[4] That is an extreme case, enormous losses under competition conditions, and your version of it will be far smaller. The direction is right even when the size is not.

What exercise does about that gap is smaller than the story suggests. The 96 kcal from lifting is genuine and it does not close a 500 kcal hole. Muscle tissue itself is usually quoted at around 13 kcal per kg per day, so two hard-won kilos buys you about 26 calories. Keep the muscle for how you move and what you can carry, not for what it burns.

What to do with all of that

Set the deficit, then stop negotiating with it

Three hundred to 500 calories below maintenance. There is a specific reason not to go deeper, and it is the most useful number in this article. Murphy and Koehler, in the Scandinavian Journal of Medicine and Science in Sports in 2022, ran a meta-regression across resistance training studies and found that a deficit of about 500 kcal a day was the point where lifting stopped adding lean mass at all.^[5] Below that you can train and hold your muscle. Above it the training stops paying you back in lean mass.

Protein at a number, not a vibe

1.6 to 2.2 g per kg of body weight. Morton and colleagues, in the British Journal of Sports Medicine in 2018, looked for the point where extra protein stops adding anything and put it at 1.62 g per kg per day, with the confidence interval running to 2.2.^[6] Two things they say themselves and most people quoting them leave out: that break point did not reach statistical significance, and the trials behind it were resistance training with protein supplements, not women eating at a deficit. So 1.6 is the number worth hitting, 2.2 is the top of the uncertainty rather than a target, and past it you are just buying expensive chicken.

Two or three sessions, and what is in them

A session is about 40 minutes and does not need to be clever. Five or six movements covering a squat, a hinge, a push, a pull and something carried, two to three sets of 6 to 12 reps, and you add a rep or a little weight whenever the last set stops feeling hard. Walking sits around it: 7,000 to 8,000 steps most days does more for your weekly energy balance than an extra gym session you dread.

There are much higher targets quoted for keeping weight off, and I want to be straight with you about where they come from. Wing and Phelan, reviewing the National Weight Control Registry in the American Journal of Clinical Nutrition in 2005, described people who had lost a lot of weight and held on to it, and they reported around an hour of activity a day.^[8] That is a self-selected, self-reported registry describing what maintainers did. It is not a trial that assigned the dose and watched it work. A reasonable direction to walk in, not a proven prescription, and treating it as one is how people end up feeling like failures at 45 minutes a day.

Where this goes wrong in practice

Eating the workout back

The version you usually hear is that everybody eats back what they burn. That is not what the best trial found. Church and colleagues, in PLoS One in 2009, randomised 411 sedentary postmenopausal women to a control group or to supervised exercise at one of three doses for six months.^[7] At 72 and 136 minutes a week, actual weight loss matched what the arithmetic predicted. Only the group at 194 minutes fell short, losing about half of the prediction. Compensation is dose-dependent, so the fix is not to distrust yourself but to notice that appetite catches up as volume climbs.

Cardio only

Cardio burns calories while you are doing it and does good things to your heart. It is not what protects your lean mass in a deficit. One of the reviews inside Bellicha's overview found resistance training cut lean mass loss by about 0.8 kg,^[2] and in the resting metabolic rate data cardio did not reach significance at all. If you have three hours a week, giving two of them to lifting will change more than giving all three to the treadmill.

Perfecting the training and ignoring the kitchen

You may have trained five days a week for months and wondered why nothing shifts. Look at the Johns numbers again: exercise-only programmes were more than 5 kg behind the combined approach at three to six months. Training hard with the food unstructured is the most exhausting way to make slow progress.

Common questions