The two popular rules and why neither holds up
The first is eight glasses of eight ounces a day, 64 fl oz. It is usually traced to a 1945 US Food and Nutrition Board recommendation of roughly 1 millilitre of water per calorie of food, about 2.5 litres for a typical adult. The next sentence of that document said most of this quantity is contained in prepared foods. The sentence was dropped, the number survived, and it has been repeated ever since without the qualification that made it true.
The second is half your body weight in ounces. For a 165 lb adult that is 82 fl oz. There is no study behind it. It is a memorable arithmetic trick that happens to land in a plausible range for people of average size and produces obvious nonsense at the extremes, since it scales with total mass rather than with metabolic rate or lean tissue. At 300 lb it asks for 150 fl oz, which is more than most people that size need and more than some can safely handle.
The formula on this page is a third rule of thumb, 30 to 33 millilitres per kilogram, and it is no better evidenced than the other two. It is used here because it at least scales with body size in a defensible way and because a number gives you something to compare against. Treat the output as an order of magnitude.
Total water is not the same as water you drink
Intake guidelines from the US National Academies and from EFSA are given as total water: everything drunk plus everything eaten. The commonly quoted US adequate intakes are about 3.7 litres a day for men and 2.7 for women, and roughly 20 percent of that typically arrives in food. Fruit, vegetables, soup, yoghurt and cooked grains all carry substantial water. Someone eating a lot of produce is drinking less than someone living on crackers and needs to.
Other drinks count as well. The idea that coffee and tea dehydrate you does not survive contact with the measurements: at habitual doses the diuretic effect of caffeine is smaller than the volume of the drink, so a cup of coffee is a net fluid gain. Alcohol genuinely does suppress vasopressin and can produce a net loss, which is why it is the exception. If you want to see how caffeine stacks up across a day, that is a separate question handled by the caffeine calculator.
The signals that actually work
Thirst is a well-calibrated regulator in healthy adults with normal access to fluid. It responds to plasma osmolality with considerable sensitivity, and the common claim that being thirsty means you are already dehydrated overstates the case: thirst rises with a change in osmolality on the order of one to two percent, which is well short of anything clinically meaningful.
Urine colour is a cruder second signal, and a useful one. Pale straw suggests adequate intake, dark amber suggests concentrating. It is confounded by B vitamins, which turn urine bright yellow regardless, by beetroot, by some medications, and by the first void of the morning, which is dark in nearly everybody. Frequency is worth noting too — most people void somewhere between four and ten times a day.
Neither signal applies to everyone. Thirst blunts with age, which is one reason older adults are over-represented in dehydration admissions. It is unreliable during hard exercise and in some neurological conditions. In those groups a schedule beats waiting for thirst.
Overdrinking is rare, real, and occasionally fatal
Exercise-associated hyponatraemia occurs when someone drinks enough plain water, over hours, to dilute blood sodium faster than the kidneys can compensate. It appears in marathons, in military training, in hot-weather manual work, and in a handful of water-drinking contests that killed the participants. Early symptoms are nausea, headache and confusion, which are also the symptoms of dehydration, so people respond by drinking more and make it worse.
The practical thresholds: the kidney can excrete somewhere around 0.8 to 1 litre per hour in a healthy adult, less in older people and less under the stress hormones released by prolonged exercise. Sustained intake above that has nowhere to go. If you are sweating for several hours, the sodium lost matters as much as the volume, and plain water alone is the wrong replacement.
What to take from the number above
Use it once, to check whether you are in the right neighbourhood. If it says 90 fl oz and you drink about that, there is nothing to change. If it says 90 and you drink 25, that gap is worth closing. If it says 90 and you already drink 150 out of a belief that more is better, the calculator is telling you to stop. Beyond that comparison, a daily tally of ounces is a chore with no demonstrated benefit for a healthy adult with a working sense of thirst.
Questions people ask
Does drinking more water improve skin, weight loss or energy?
Correcting actual dehydration helps all three, because dehydration impairs cognition, makes skin less turgid and is sometimes mistaken for hunger. Going beyond adequate intake has not been shown to do anything further. Trials of extra water for skin appearance are small and inconsistent; trials for weight loss show a modest effect that is best explained by water displacing calorie-containing drinks rather than by hydration itself. If you replace two sodas a day with water you will probably lose weight, and the mechanism is the sodas.
Do coffee and tea count toward my intake?
Yes. The diuretic effect of caffeine at habitual intakes is smaller than the fluid volume of the drink, so a mug of coffee is a net gain of water. This has been measured directly, including a trial comparing coffee against equal volumes of water in habitual drinkers, which found no difference in hydration status. The effect is larger in someone with no caffeine tolerance drinking a great deal at once, but even then it does not reverse the sign. Alcohol is the genuine exception: it suppresses vasopressin and beer or spirits can produce a net fluid loss.
Is a set amount of water useful for kidney stones?
This is the one situation where a fluid target has real trial evidence behind it. For people with a history of calcium stones, a urine output of about 2 to 2.5 litres a day cuts the recurrence rate substantially, which typically means drinking well over the figure this calculator produces. That target belongs with a clinician who knows your stone type and your kidney function, not with a general formula. If you have never had a stone, the preventive case for extra water in the general population is much weaker.
How much should children drink?
This calculator is for adults and the per-kilogram figure will overestimate at low body weights. The paediatric maintenance rule in common clinical use is the Holliday-Segar method: 100 ml per kg for the first 10 kg of body weight, 50 ml per kg for the next 10 kg, and 20 ml per kg beyond that. For a 25 kg child that is 1,600 ml a day of total water, food included, which is a good deal less than a body-weight rule would suggest. Children also dehydrate faster in heat and are worse at noticing it, so on hot days offer drinks rather than waiting to be asked.
Should the number change when I am ill?
Fever, vomiting and diarrhoea all increase losses, and diarrhoea in particular loses electrolytes as well as water, which is why oral rehydration solutions exist and why plain water is not the ideal replacement for it. Going the other way, some conditions require less: heart failure, advanced kidney disease and cirrhosis all commonly come with a prescribed fluid restriction, and in those cases exceeding it is the harm. Any illness serious enough to make you ask the question is serious enough to ask a clinician instead of a calculator.