What the rule says
The 3,500-calorie rule turns weight loss into simple division. Cut 500 kcal a day and you save 3,500 kcal a week, so you lose one pound a week. Cut 1,000 kcal a day and you lose two. In metric terms the rule is about 7,716 kcal per kg, which Kevin Hall of the NIH writes as 32.2 MJ per kg.
The rule is everywhere. Thomas and colleagues wrote in 2013 that it still appears in textbooks, on respected government and health websites, and in scientific papers. The goal timeline in the calorie calculator on this site uses it as well, which is why this guide explains what that date means and what it leaves out.
The key feature of the formula is that time only multiplies. Double the days and you double the loss. Nothing in it changes as you get lighter.
Where the 3,500 comes from
The figure is usually traced to a 1958 paper by M. Wishnofsky in the American Journal of Clinical Nutrition, titled Caloric equivalents of gained or lost weight. Thomas and colleagues name it as the source of the rule, and Hall and Chow cite it for the calculation behind it: a pound of adipose tissue, or body fat tissue, stores about 3,500 kcal. In a 2008 paper Hall describes the assumption more precisely. The derivation assumes that everything you lose is fat tissue, and that this tissue is 87 percent fat.
As a description of the energy stored in a pound of fat tissue, that is a reasonable estimate. The trouble starts when it is used to predict how a person loses weight over months and years.
Why it overestimates long-term loss
Your burn falls as your weight falls
This is the big one. The 2011 Lancet paper by Hall and colleagues says the rule ignores the physiological adaptations to a lower body weight, which change both your resting metabolic rate and the energy cost of physical activity. A lighter body needs less fuel to run and less to move around.
You can see the direction of this in the formula behind the calorie calculator. In the Mifflin-St Jeor equation each kilogram of body weight adds 10 kcal to your estimated BMR, and at the moderately active factor of 1.55 that becomes 15.5 kcal of daily burn per kilogram. So after losing 10 kg, the same menu that started as a 500 kcal deficit is only about 345 kcal below your new needs. The BMR vs TDEE guide works through that formula. The Lancet model tracks these changes in much more detail, but the direction is the same: the deficit shrinks as you lose.
Hall and Chow call this failure to account for changes in energy balance over time the most serious error of the rule. A shrinking deficit means the loss slows, and when intake and burn meet again, it stops.
Not every pound you lose is fat
Weight loss also takes some lean tissue, which stores far less energy. The Lancet paper puts the energy content at 39.5 MJ per kg of body fat, about 9,441 kcal, and 7.6 MJ per kg of lean mass, about 1,816 kcal. The mix decides how many calories a pound of loss really costs.
Hall's 2008 paper found that this energy density depends on how much fat you start with. The 3,500 figure roughly matched people with more than 30 kg of body fat at the start. For people with less than about 25 kg of fat it was too high, so they needed a smaller deficit per kilogram lost. The same paper notes that larger losses tend to come with a lower average deficit per kilogram, because they contain more lean mass. So this part of the error can push either way. Hall and Chow add that swapping in a better number for the tissue would not repair the rule, because the time effect above still remains.
Real life adds slippage
The Lancet model assumes you stick to the change perfectly and does not automatically simulate any compensation. Its authors write that limited adherence and compensatory effects are likely to play a prominent role. So even the dynamic model describes the best case for a given cut, not a guarantee.
What the NIH model predicts instead
In 2011 Hall and colleagues published a mathematical model of how body weight responds to a change in energy intake. It is the model behind the NIH Body Weight Planner, run by the National Institute of Diabetes and Digestive and Kidney Diseases, which asks anyone who publishes results from the planner to cite this paper. The paper also offers a short rule of thumb for an average overweight adult:
In the paper's own metric units, every permanent change of 100 kJ a day, about 24 kcal, leads to an eventual change of about 1 kg. The paper describes the response as slow, with half times of about a year, and notes that adults with more body fat can expect a larger loss for the same change in intake.
The difference from the static rule is the word eventual. A lasting cut does not keep taking weight off at the same pace. It moves you toward a new, lower weight where your smaller body burns what you now eat. Writing about the planner in 2018, Hall put it plainly: the body responds in a way that offsets the deficit over time, so people lose much less weight than the 3,500 kcal rule promises.
Worked example: the man from the Lancet paper
The Lancet paper simulates a sedentary man who weighs 100 kg (220 lb) and cuts his intake by 2 MJ, or 480 kcal, a day for good. Here is what each method says.
| Time | Static 3,500 kcal rule | Dynamic model (Hall et al. 2011) |
|---|---|---|
| After 1 year | 50 lb lost (23 kg) | About half of the eventual loss, roughly 11 to 13 kg (about 25 lb) |
| After 3 years | 150 lb lost (68 kg), leaving him at 70 lb | About 95% of the eventual loss, roughly 24 kg (52 lb) |
| After 10 years | 501 lb lost, more than twice his starting weight | A plateau at about 75 kg (165 lb), 25 kg (55 lb) below the start |
The arithmetic for the static column is simple: 480 × 365 = 175,200 kcal a year, divided by 3,500 is about 50 lb. The paper gives 22 kg and says this first-year figure is about 100% greater than its model's prediction. Its model, by contrast, has the man settle near 75 kg, a loss of about 25 kg, and reach half of that in roughly a year.
In the first year the static rule is off by about a factor of two. After that it is not just optimistic but impossible: by year ten it has the man losing more than he weighs. The model's curve bends and flattens, because each kilogram he loses lowers his daily burn and shrinks the gap the 480 kcal cut created.
The two rules at common deficits
The table puts the static rule next to the Lancet rule of thumb for an average overweight adult. Both assume the change is kept exactly, every day.
| Lasting daily cut | Static rule, 1 year | Static rule, 3 years | Model, about 1 year | Model, about 3 years | Model, eventual |
|---|---|---|---|---|---|
| 100 kcal | 10.4 lb | 31.3 lb | 5.0 lb | 9.5 lb | 10.0 lb |
| 250 kcal | 26.1 lb | 78.2 lb | 12.5 lb | 23.8 lb | 25.0 lb |
| 500 kcal | 52.1 lb | 156.4 lb | 25.0 lb | 47.5 lb | 50.0 lb |
| 750 kcal | 78.2 lb | 234.6 lb | 37.5 lb | 71.3 lb | 75.0 lb |
Two patterns stand out. After one year the static rule predicts about twice what the model does, the same gap the paper reports for its worked example. And after three years the static rule keeps climbing in a straight line while the model is already close to its end point. For the 500 kcal row, the classic pound a week, the rule says 156 lb in three years and the model says about 48.
The rule of thumb is an average. The paper's 100 kg man loses about 25 kg from a 2 MJ cut, a little more than the roughly 20 kg the average rule gives, because he starts with more body fat. Individual results also spread out, and the paper notes that imperfect adherence widens that spread further.
What real studies found
Thomas and colleagues tested the static rule against seven weight loss experiments in which people were either kept in confinement under full supervision or had their intake measured objectively. Across 103 adults, 71 men and 32 women, the average actual loss was 20.1 lb. The 3,500 kcal rule had predicted 27.6 lb. In the five studies where they could also run the dynamic model, it predicted 18.4 lb against an actual 17.0 lb. Their conclusion: the 3,500 kcal rule grossly overestimates actual weight loss, and dynamic models are a more accurate alternative.
They also point out that beyond one year the gap between the two predictions becomes substantial, because the model produces a curved path that levels off while the rule draws a straight line.
Using this with the calculators
- Read a 3,500-based timeline as the optimistic case. The calorie calculator's goal date assumes the deficit stays the same size. The longer the timeline, the further it runs ahead of what the model predicts.
- Recalculate as you go. Put your new weight into the calorie calculator after a noticeable change and take any deficit from the fresh number. The macro calculator then splits the new total into protein, carbs and fat.
- Treat exercise calories as estimates. The calories burned calculator uses MET values from the 2024 Adult Compendium of Physical Activities, whose authors write that it does not reflect precise individual energy expenditure. A burn figure from a workout is a starting point, not a deposit you can bank at 3,500 kcal per pound.
- Expect a plateau. In the model, leveling off is the normal end of a fixed cut, not a sign that something went wrong. Hall describes the Body Weight Planner as a tool that shows what people have to do permanently to keep the weight off.
If you want a projection that curves the way the research says it should, the NIH Body Weight Planner runs the dynamic model for you. Its page says it is for adults 18 and older and not for pregnant or breastfeeding women. If you are under 18, pregnant, ill, or take medication that affects your weight, talk to a doctor or dietitian before you change how much you eat.
Frequently asked questions
Is a pound of fat really 3,500 calories?
Roughly, for fat tissue. Kevin Hall and Carson Chow write that the rule came from calculating that a pound of adipose tissue stores about 3,500 kcal, and Hall traces the calculation to an assumption that the lost tissue is 87 percent fat. The catch is that not every pound you lose is fat tissue, and, more importantly, that your daily burn falls as you get lighter. Hall and Chow add that a corrected number for the tissue would not repair the rule, because the bigger error is ignoring how energy balance changes over time.
Does a 500-calorie daily deficit lose a pound a week?
That is what the static rule predicts, and it predicts it forever. The rule of thumb from the 2011 Lancet model points to an eventual change of about 50 lb for a lasting 500 kcal cut in an average overweight adult, with about half of it, roughly 25 lb, reached after a year. That averages about half a pound a week over the first year, slowing after that, and it assumes you keep the cut exactly.
Why does weight loss slow down or stop?
Because a smaller body burns less. The 2011 Lancet paper says the 3,500 kcal rule ignores the physiological adaptations to a lower weight that change both the resting metabolic rate and the energy cost of physical activity. As your burn drops, the same intake leaves a smaller and smaller gap, until intake and burn meet at a new, lower weight. In the paper's example that happens near 75 kg for a man who started at 100 kg.
What is the NIH Body Weight Planner?
A free online tool from the National Institute of Diabetes and Digestive and Kidney Diseases that lets you plan calories and physical activity to reach a goal weight in a set time and keep it there. NIDDK asks people who publish results from it to cite the 2011 Lancet paper by Hall and colleagues, the dynamic model this guide describes. Its page says it is for adults 18 and older, not for pregnant or breastfeeding women, and is not intended to provide medical advice.
Has anyone tested the 3,500 kcal rule against real weight loss?
Yes. Thomas and colleagues compared it with seven weight loss experiments in which intake was either supervised in confinement or measured objectively, 103 adults in all. On average people lost 20.1 lb, while the rule predicted 27.6 lb. The authors write that the rule grossly overestimates actual weight loss and offer tools based on a dynamic model instead.
Does the rule work the same way for weight gain?
The Lancet rule of thumb is stated for any permanent change of intake, up or down: about 10 kcal a day for each pound of eventual weight change. So the same logic applies to gain. A lasting extra 100 kcal a day points to roughly 10 lb more in the long run, not the 10 lb a year, year after year, that the static rule would add up.
Sources
- Wishnofsky M, Caloric equivalents of gained or lost weight, American Journal of Clinical Nutrition 6(5):542 to 546, 1958 (citation record)
- Hall KD, Sacks G, Chandramohan D, Chow CC, Wang YC, Gortmaker SL, Swinburn BA, Quantification of the effect of energy imbalance on bodyweight, The Lancet 378(9793):826 to 837, 2011
- Hall KD, What is the required energy deficit per unit weight loss?, International Journal of Obesity 32(3):573 to 576, 2008
- Thomas DM, Martin CK, Lettieri S and others, Can a weight loss of one pound a week be achieved with a 3500-kcal deficit? Commentary on a commonly accepted rule, International Journal of Obesity 37:1611 to 1613, 2013
- Hall KD, Chow CC, Why is the 3500 kcal per pound weight loss rule wrong?, International Journal of Obesity 37(12), 2013
- Mifflin MD, St Jeor ST and others, A new predictive equation for resting energy expenditure in healthy individuals, American Journal of Clinical Nutrition, 1990
- NIDDK, Body Weight Planner
- NIDDK, Body Weight Planner (research page with the paper to cite and the disclaimer)
- NIDDK, Diabetes Discoveries and Practice blog: NIH Body Weight Planner, Kevin Hall, July 11, 2018
- Herrmann SD and others, 2024 Adult Compendium of Physical Activities, Journal of Sport and Health Science 13(1):6 to 12, 2024
- NIST Special Publication 811, Guide for the Use of the SI, Appendix B.9 (calorie to joule)
General information based on the sources above, not medical advice.