On a one-hour run, a bathroom scale that reads to 0.1 kg can put a sweat rate up to 0.1 litre an hour out through rounding alone; in our simulation of 200,000 sessions, one in four landed at least 0.05 L/h off. The rounding's standard deviation, on our arithmetic, is about 3–4% of a typical hourly rate, smaller than the 5–7% by which one adult's rate varies from day to day in Baker's 2017 review. The measurement is weight before, minus weight after, plus fluid drunk, minus any urine, divided by the hours, with a kilogram read as a litre. Measured that way, 36 Singapore Armed Forces regulars racing in Singapore in 2023 averaged an estimated 1.3 litres an hour.
Every protocol measures sweat rate as mass lost plus fluid drunk
The National Athletic Trainers' Association (NATA) gives the formula in Table 2 of its 2017 position statement, which works "where mass = volume", so the answer comes out in litres an hour. By our arithmetic, a runner who goes from 70.5 to 69.0 kg in an hour while drinking 500 ml has sweated about 2.0 litres.
Baker's review, written at PepsiCo's Gatorade Sports Science Institute, uses the same subtraction, counts food eaten as intake and asks for every weighing to be done twice. The protocols disagree on clothing. Baker prefers nude weighing; NATA wants tests repeated "several times under realistic environmental conditions" in the sport's own kit. The Korey Stringer Institute's sweat-rate sheet says "nude is best", or else the same clothes both times. Daily water needs are covered in our guide to the eight-glasses rule.
Runners in Singapore and Thailand averaged inside Baker's typical range
Two tropical field studies measured how much runners sweat in humid heat, and found ordinary rates. Sakamoto and colleagues, funded by Singapore's Ministry of Defence, weighed Singapore Armed Forces regulars nude at the 2023 SAFRA Singapore Bay Run and Army Half Marathon, in 87% relative humidity. The 36 in its fluid-balance group averaged 1.3 ± 0.4 L/h; the paper's medians are 1.2 for the 21-km race and 1.4 for the 10-km. Urine was not counted, which pushes the estimates up.
In Thailand, Surapongchai and colleagues at Mahidol University weighed 166 recreational runners, clothed, on light-to-moderate practice runs at 29.6 °C. Men averaged 1.3 L/h and women 0.9, individually 0.1 to 3.8 L/h in its table, a wider span than its discussion quotes. Sweat held in clothing pulls those figures down. The study, funded by Suntory PepsiCo Beverage Thailand and the Gatorade Sports Science Institute, concluded that current fluid guidelines "are applicable to tropical native runners". Its 70% humidity is hard to set beside Singapore's 87%, since relative humidity depends on temperature.
Heat still counts within one person. Baker lists heat acclimation as raising sweating rate, and the Korey Stringer Institute writes that rates generally rise after 10 to 14 days of heat exposure. A figure taken on a cool morning describes a cool morning.
Rounding on the scale registers most in a half-hour test
We modelled rounding alone, for runners of 50 to 90 kg. Two readings each rounded to 0.1 kg can stack to a full 0.1 kg of error, and a shorter session divides it by fewer hours.
| Session on a 0.1-kg scale | Worst rounding error | Sessions off by 0.05 L/h or more |
|---|---|---|
| 60 minutes | ±0.10 L/h | 25% |
| 45 minutes | ±0.13 L/h | 39% |
| 30 minutes | ±0.20 L/h | 56% |
The Korey Stringer Institute's 30-minute test protocol, which multiplies the result by two to get an hourly rate, doubles the rounding error along with the sweat loss. The Thai study read to 0.10 kg over runs of about 44 minutes, up to about ±0.14 L/h per runner on our arithmetic; Sakamoto's study read to 0.02 kg. For a 70-kg runner, the same 0.1 kg moves the percentage loss by up to 0.14 of a point across the two weigh-ins. Our model covers resolution only; we found no source for how accurate consumer scales are.
The scale weighs breath and burned fuel as sweat
Baker puts fuel burned and water breathed out at "approximately 5–15% combined" of the mass lost, a correction that matters most in long, hard or dry sessions. That error always inflates the sweat figure. No source we opened sizes it for a humid one-hour run.
In a study of women running 30 km by Cheuvront and colleagues, which Baker cites, ignoring breath and fuel overstated sweat loss by 9% in a warm trial and 20% in a cool one. Ignoring urine overstated it by 16% and 37%, while sweat trapped in clothing cut it by 8–10%.
ACSM gauges its two per cent goal against performance
The American College of Sports Medicine's 2007 position stand says in its abstract that drinking during exercise aims to prevent "excessive (>2% body weight loss from water deficit) dehydration" and so "avert compromised performance". It adds that "Individual sweat rates can be estimated by measuring body weight before and after exercise."
NATA's concern with the 2% threshold goes beyond performance. For high-performance athletes, staying within it "helps to maintain the physiological, perceptual, and safety aspects of their exercise". Its recommendation 4, graded B on its strength-of-recommendation (SOR) scale, puts compromised temperature control at hypohydration, a body-water deficit, "greater than 1%", and rising heat-illness risk above 3%. For athletes who do not know their sweat rates, NATA writes that "drinking to thirst during activity most likely represents a safe strategy to prevent overdrinking". Twelve elite runners at the 2017 Standard Chartered Singapore Marathon lost 4.6% on average and, their authors report, performed "without medical complication"; that describes those twelve and nobody else.
Our sweat rate calculator grades a session against these lines. Until 8 October 2026 the calculator labelled every weight gain "Hydration excellent", all 880 gains in our test, which was wrong: NATA advises against any gain. Below 0.4 L/h, its recommended intake also exceeded the sweat lost. The version fixed that day treats a gain as its own band, sets no drinking target and shows the rate to 0.1 L/h with its rounding margin.
Some tropical marathon and ultramarathon runners tipped the scales heavier at the finish
NATA's recommendation 5, graded A, ties exercise-associated hyponatraemia, low blood sodium, to drinking too much hypotonic fluid, "resulting in maintained or increased body weight during exercise lasting 1 hour or more". Recommendation 6, also A, says individuals "should not gain body mass (or body weight) during exercise" unless they begin "with an unavoidable fluid deficit". NATA lists "acute weight gain" among the signs, and it takes a blood sodium test to tell the condition apart.
Lee and colleagues, whose first author was at DSO National Laboratories in Singapore, weighed runners at a 42-km and an 84-km race held at 29.0 °C. In the marathon 22 runners (7%) gained weight, and in the longer race 9 (8%), about one in fourteen across those two fields. Three of the eight runners with symptoms had hyponatraemia, in a paper its authors titled "First reported cases of exercise-associated hyponatremia in Asia".
At the 2023 Singapore races, Sakamoto's runners replaced about half their sweat loss on average, from none of it to 277%. One runner in the 10-km race finished 0.52% heavier than at the start.
View author profile → · Editorial policy
