The people who keep weight off long term are studied less than the people who lose it, which is backwards, because maintenance is the part that fails. The best available dataset on this is the National Weight Control Registry, started in 1994 by Rena Wing at Brown and James Hill at Colorado, which tracks people who have maintained a loss of at least 30 pounds for at least a year.
Its members have lost an average of 66 pounds and kept it off for an average of 5.5 years. The spread is enormous: losses from 30 to 300 pounds, maintenance durations from 1 year to 66 years. About 80 percent are women, and the average woman in the registry is 45 years old.
what they share, and what they do not
The diet is where people expect the answer to be, and it is where the registry is least informative. Members got there by many different routes. Roughly half lost weight through a program and half on their own. Almost all of them report modifying food intake in some way, but the specific approach varies so widely that no single diet emerges as the method.
The behavioral patterns are much more consistent. Published analyses of the registry report that around 90 percent exercise on average about an hour a day, roughly 78 percent eat breakfast daily, about 75 percent weigh themselves at least once a week, and about 62 percent watch fewer than 10 hours of television a week.
That hour a day of activity is the number people flinch at, and it is worth stating clearly rather than softening. It is far above the 150 minutes a week in the Physical Activity Guidelines for Americans. Most of it, in registry data, is walking. Not an hour of intense training, an hour of movement.
the honest limitation of this data
Every one of those figures comes from a self-selected, self-reported sample of people who already succeeded. That is a serious methodological problem and anyone citing the registry without mentioning it is misleading you.
The registry cannot tell you what causes maintenance, only what maintainers report doing. People who joined did so because they succeeded, and we have no comparison group of people who did the same things and failed. Self-reported food intake is systematically underestimated across the whole nutrition literature, often by 20 percent or more. The breakfast finding in particular has not held up in randomized trials: when breakfast eating is actually assigned, it does not reliably produce weight loss.
So treat the registry as a description of a population, not a prescription. The daily weighing and the high activity volume have support from other study designs, which makes them more believable. The breakfast number is probably a correlation with something else, like having a structured day.
the biology that makes this hard
Maintenance is difficult for reasons that have nothing to do with willpower, and the research on this is unkind.
Weight loss reduces energy expenditure by more than the change in body size predicts, a phenomenon called adaptive thermogenesis, and in some studies it persists for years. Leptin falls, ghrelin rises, and appetite increases in a way that is measurable and does not simply fade. A 2016 follow-up study of Biggest Loser contestants found resting metabolic rate still suppressed six years after the competition in participants who had regained most of their weight.
That means a woman maintaining a 60-pound loss is eating less than a woman of the same current weight who was never heavier, and she is hungrier while doing it. Which reframes the hour of daily activity from an admirable habit into an adaptation to a real physiological constraint: it partly offsets the expenditure gap.
the practices with the best supporting evidence
| Practice | Evidence quality |
|---|---|
| Regular self-weighing | Good, supported by randomized trials as well as registry data |
| High daily activity volume | Good for maintenance specifically, weaker for initial loss |
| Adequate protein intake | Good for appetite and for preserving lean mass during loss |
| Consistent routine including weekends | Moderate, registry members report less weekday/weekend variation |
| Eating breakfast | Weak when tested experimentally |
| Specific diet composition | No single approach outperforms long term |
The self-weighing finding deserves emphasis because it is counterintuitive and it is also genuinely contraindicated for some people. Frequent weighing helps most maintainers catch a small regain before it becomes a large one, which is the entire mechanism. For anyone with a history of disordered eating, the same behavior can be harmful, and measuring progress by waistband fit or training performance is a reasonable substitute.
the number that matters more than the target
Registry data suggests maintenance gets easier with time, with the risk of regain dropping substantially after two to five years of successful maintenance. That is one of the more encouraging findings in the dataset, and it reframes the first two years as the period that requires the most deliberate structure.
A 5 to 10 percent weight loss maintained produces most of the measurable metabolic benefit: improved blood pressure, better insulin sensitivity, improved lipids. For a 200-pound woman that is 10 to 20 pounds, not 66. The registry’s average of 66 pounds reflects who chooses to join a weight loss registry, not the threshold at which health outcomes improve.
Anyone with a metabolic condition, a history of eating disorders, or who takes medication affecting weight should be having this conversation with a physician or a Registered Dietitian Nutritionist rather than reading population statistics. The registry describes averages, and averages are not advice.