What actually works in exercise

Fitness advice arrives at a volume and confidence entirely out of proportion to how much genuinely changes. New methods appear constantly; the underlying physiology has not moved in decades. What follows is the part that has survived repeated scrutiny, which is less exciting than the marketing and considerably more useful. None of it is personal medical advice, and anything specific to your own health belongs with a doctor.

Adaptation is the whole mechanism

Exercise does not build anything directly. It creates a stress the body is not currently equipped for, and the building happens afterwards, during recovery, as the body adjusts to handle that stress more comfortably next time.

Everything else follows from this. If the stress never increases, adaptation stops, because the body is already equal to the demand. If the stress increases faster than recovery can keep up, you accumulate fatigue and injury rather than fitness. Training is the management of that gap.

It also explains why rest days are not time off from training. They are when training takes effect. Someone who trains hard seven days a week is not training more than someone doing four; they are adapting less from the same work.

Progressive overload, which is unglamorous and non-negotiable

The single most reliable principle is that the demand has to keep rising. More weight, more repetitions, more distance, less rest between efforts, better control of the movement — any of these count.

What does not work is repeating the same session indefinitely and expecting continued change. This is the most common reason people plateau after a promising start: the first months bring rapid improvement because almost anything is a novel stress, and then the same routine stops being a stress at all.

The corollary is that progress needs measuring. Not obsessively, but enough to know whether the demand is actually increasing, because memory is unreliable and "it felt hard" is not a measurement.

Strength and cardio do different jobs

Resistance training and endurance work adapt different systems, and neither substitutes for the other.

Loading muscle produces small amounts of damage and mechanical stress, and the repair response increases the muscle's ability to produce force. The adaptations are local to what you trained and specific to how you trained it. Crucially, resistance training also loads bone, and bone responds by becoming denser — one of the few interventions that meaningfully affects bone density, and increasingly relevant with age.

Sustained aerobic work adapts a different set of systems: the heart pumps more blood per beat, the vasculature improves, muscles build more mitochondria and get better at using oxygen. Cardiorespiratory fitness is among the strongest predictors of long-term health outcomes in the epidemiological literature — stronger than most of the variables people worry about more.

The practical answer for general health is not to choose. It is to do some of each, which is what essentially every public health guideline recommends and what almost every training argument online manages to obscure.

The dose is smaller than people fear

A recurring finding is that the largest health returns come from the first increments of activity, not the last. The gap between doing nothing and doing a modest amount is far larger than the gap between a moderate amount and a great deal.

This is genuinely encouraging news, and it is buried by a culture that presents training as an identity requiring substantial time. Standard guidance lands around 150 minutes of moderate activity a week plus two sessions involving resistance, and that is a target, not a threshold — being under it is enormously better than being at zero.

What the evidence does not support

Several widely repeated claims hold up poorly, and it is worth naming them.

Spot reduction — exercising a body part to lose fat specifically from it — does not work. Fat is mobilised systemically, and where it comes from is largely a matter of individual physiology. Abdominal exercises build abdominal muscle and do not preferentially remove abdominal fat.

Static stretching before exercise as injury prevention has weaker support than its ubiquity suggests, and stretching held for long periods immediately before may temporarily reduce force output. A general warm-up that raises temperature and rehearses the movement has better evidence behind it.

Soreness is not a measure of quality. Delayed onset muscle soreness relates to unfamiliar movement more than to effective training, and it fades as you adapt to an exercise while the training keeps working. Chasing soreness optimises for novelty rather than progress.

And exercise is a comparatively weak tool for weight loss on its own. It is excellent for health, function and maintaining muscle; the energy expenditure is smaller than most people estimate and is partly offset by compensatory changes. This is not an argument against exercising. It is an argument against judging it by the wrong outcome.

The part that actually decides it

Almost every comparison between reasonable training methods finds smaller differences than the arguments about them would suggest. Which means the variable that dominates is not programme design but whether the programme happens at all, repeatedly, for years.

Adherence is the real limiting factor, and it is a design consideration rather than a character trait. A slightly suboptimal routine you will keep doing beats an optimal one you abandon in March. Convenience, enjoyment and a schedule that survives a bad week are legitimate training variables, and treating them as such is not a compromise — it is the intervention with the largest effect size available.