Everything you know about dieting is wrong, say United States scientists, who have devised a formula for calculating calories and weight loss they hope will change the way people tackle obesity.
Obesity rates have doubled worldwide in the past 30 years, coinciding with a growing food surplus.
The ensuing epidemic has sparked a multibillion-dollar weight-loss industry that has largely failed to curb the problem.
Current standards advise people that cutting calories by a certain amount will result in weight loss over time.
But that advice fails to account for how the body changes as it slims down, burning less energy and acquiring a slower metabolism, researchers told an American Association for the Advancement of Science meeting in Vancouver.
The result is a plateau effect that ends up sending dieters back into harmful patterns of overeating.
As an example, researcher Kevin Hall offered up his large vanilla latte, purchased at a popular coffee shop. When he asked, the barista told him it contained about 240 calories.
"The notion was if I drank one of these every day and then I replaced it with just black coffee no sugar, then over the course of a year I should lose about 25 pounds [11.5kg], and that should just keep going," Hall said.
"People have used this sort of rule of thumb to predict how much people should lose for decades now, and it turns out to be completely wrong."
Hall, a scientist with the US National Institutes of Health, said: "If I want to lose 10 pounds of weight eventually, I have to cut 100 calories per day out of my diet.
"You'll get halfway there in about a year, and then you will eventually plateau, [reaching the goal] after about three years. The contrast is the old rule of thumb predicts twice as much weight loss after a year, and it gets worse after that."
The new model gives dieters one calorie goal for short-term weight loss and another for permanent weight loss.
Tests on small numbers of adults who were fed strictly controlled diets showed the model was accurate, though real-life situations are harder to predict.
Study co-author Carson Chow said the daily calorie cut needed for weight loss was actually smaller than researchers anticipated.
"It is essentially one cookie different a day, so a 150-calorie cookie leads to a seven kilogram difference in weight. That is huge in my opinion," Chow said.
"People can plug in some information about their initial age, their height, their weight, some estimate of their physical activity level," Hall said.
Add in a goal weight and the "model will simulate what changes of diet or exercise that person would have to do to achieve that goal weight, and then more importantly, what they need to do to permanently maintain that weight loss".
Since the The Lancet article appeared, the notion has not exactly taken the world by storm, in part because it is not primed for public use, but is mainly aimed at doctors and researchers with adult patients. Also, if a dieter enters an extreme weight goal, the number of calories the model returns may be much too low to be realistic or healthy, so it needs an expert's interpretation.
"It's not particularly user-friendly ... but it is still relatively informative," said Hall, who maintains hope that some day his message will be heard.