Caffeine and performance: how much, when and for whom

You have a double espresso before training, the top set feels smoother, and you honestly cannot tell whether that was the coffee or just a good day. Reasonable question — and the follow-up matters more: how much do you actually need, when should you take it, and does any of it apply to someone lifting three evenings a week after work rather than racing on a Sunday morning?

Caffeine is one of the few supplements where the evidence base is genuinely thick. The catch is that it works does not mean it works equally for everyone, at every dose, for every kind of training. So here is the split between what is reasonably settled and what merely gets repeated a lot.

What we know reliably

Endurance improves. This is where the evidence is most consistent. For efforts lasting from a few minutes to a few hours — running, cycling, rowing, long intervals — caffeine typically buys something in the region of two to four percent. It looks small written down, but you feel it in the back half of a hard session.

Strength and reps improve, but less. The findings here are patchier than for endurance, though they mostly point the same way: slightly higher maximal strength, and usually one or two extra reps in a set taken to failure. The effect shows up more clearly in higher-rep work than in a single all-out attempt.

The same work feels easier. This is one of the steadiest findings in the whole area: at a matched load, people on caffeine report lower perceived effort and less muscular discomfort during the set. A good chunk of the performance gain probably comes from exactly that.

The best-studied dose is 3 to 6 mg per kilogram of body weight. For an 80 kg person that is roughly 240 to 480 mg. What matters is that the bottom of that range performs almost as well as the top, with noticeably fewer side effects.

Timing: 45 to 60 minutes before the work. Blood levels usually peak somewhere between thirty and ninety minutes after you take it, so that window is dependable enough to plan around.

It wrecks sleep, more than people admit. The half-life is roughly four to six hours, longer in some people. A dose taken at five in the afternoon is still doing something at midnight for plenty of us. The evidence that caffeine within several hours of bedtime shortens sleep and degrades its quality is strong — including in people convinced it does not affect them.

It does not dehydrate you. At moderate doses the diuretic effect is small and practically irrelevant, especially in regular consumers. Coffee counts towards your daily fluids.

What is probably true, but not settled

Tolerance. Habitual users almost certainly still benefit from an acute dose — the idea that coffee stops working if you drink it daily has not held up. It is possible the effect is somewhat blunted in them. That question is not resolved.

Withdrawal before a competition. Cutting caffeine for seven to ten days so it hits harder later is a popular tactic on thin, inconsistent evidence. The cost is certain — headaches, fatigue, poor sessions during the washout — while the payoff is not.

Genetics. There are gene variants that influence how fast you clear caffeine and how jittery it makes you. The logic is appealing, but results point both ways. Tests promising you a precise personal dose are selling more certainty than the science currently has.

Format. Capsules, coffee and caffeinated gum all work; gum absorbs fastest. Coffee's only real drawback is dose variability — the same cup can carry anywhere from about seventy to over 150 mg depending on the beans and the brew. Mouth-rinsing with a caffeine solution is thinly researched so far.

Women are under-studied. Most of the work has been done in men. There is no strong reason to think caffeine behaves differently in women, but questions like the influence of cycle phase or oral contraceptives on the response sit firmly in the under-researched pile.

Pairing it with creatine. The old claim that caffeine cancels creatine came out of a very narrow context and is now considered weakly supported. If you want the separation of proven from marketed on that one, we covered it in a dedicated piece on creatine.

What gets garbled in the retelling

That more is always better. Above roughly 9 mg per kilogram there is no extra benefit — only tremor, a racing pulse, anxiety and a distressed stomach. The benefit curve flattens long before the side-effect curve does.

That the sensation tells you it is working. The skin tingle from a pre-workout comes from an ingredient with nothing to do with the caffeine, and a fast heart rate is not a performance measure. Caffeine can work without a single dramatic sensation.

That pre-workouts work because of the whole blend. In most such products, the bulk of what you feel is the caffeine. The other ingredients range from decently studied to barely studied.

That caffeine burns fat. It nudges energy expenditure up acutely, but that does not translate into meaningful fat loss without a calorie deficit. As a weight-loss tool it is close to useless.

That it substitutes for sleep. It masks fatigue rather than removing it. Technique and decision-making still suffer, and chronic short sleep is only fixed by sleeping.

What this means for your training and diet

If you want to try it, start at 3 mg per kilogram — around 200 mg at 70 kg, 240 mg at 80 kg — taken 45 minutes out. If that sits well and you want to see whether there is more on the table, move to 4 or 5 mg/kg, but never for the first time before an important session or a competition.

Add up every source: coffee, tea, energy drinks, pre-workout, chocolate. For healthy adults, the ceiling most commonly cited is about 400 mg a day.

If you train after five in the afternoon, seriously consider skipping it or halving the dose. A few percent on your reps is not worth ninety minutes of lost sleep — recovery and adaptation happen overnight, not in the set.

And the part people skip: caffeine is the last five percent, not the first fifty. If you are not hitting enough weekly sets per muscle group, or your rest between sets does not match what you are training for, caffeine will not paper over it. Past fifty there is an extra reason for restraint, since sleep is more fragile anyway — worth reading how sensible training at fifty actually looks.

When to see a doctor

Check with a doctor before using it regularly if you have high blood pressure, a heart rhythm disorder, an anxiety disorder, reflux problems, or if you take medication — caffeine interacts with several drug classes. In pregnancy the usual advice is a limit of around 200 mg a day. If a dose leaves you with an irregular or very fast heartbeat, chest pain or faintness, stop and get it looked at.

Short version

• It works, and that is not in dispute: strongest for endurance, smaller but real for strength and reps.
• 3 to 6 mg/kg, 45 to 60 minutes before. The lower end performs nearly as well with fewer side effects.
• Training feeling easier is part of the mechanism, not an illusion.
• Sleep is the main cost. Late-afternoon and evening sessions are a reason for a smaller dose or none.
• Tolerance, genetics and pre-competition withdrawal are open questions, not rules.
• Do not spend it on a badly planned programme and too little sleep.