Most of them. A 2018 review by Kevin Hall and Scott Kahan in Medical Clinics of North America, looking at the maintenance of lost weight and the long-term management of obesity, estimates that roughly 80% of people who lose weight return to or exceed their starting weight within three to five years. The mirror image of that figure comes from Rena Wing and Suzanne Phelan's 2005 review in the American Journal of Clinical Nutrition, which put successful long-term maintenance at about 20% of people who lose weight.
The short answer
Roughly four in five regain within three to five years. Roughly one in five keep at least a tenth of their body weight off for a year or more. Those two figures describe the same distribution from opposite ends, and neither is a statement about willpower, since they hold across very different intervention types.
Where the 80% figure comes from
Hall and Kahan's review synthesises a broad body of intervention data rather than running a new study, and the pattern it describes is remarkably consistent across intervention types: an early period of weight loss, then a plateau, then progressive regain. That shape holds whether the intervention was dietary, behavioural, or a combination.
The consistency is the interesting part. If regain were mostly about which method someone picked, you would expect the curves to differ meaningfully between methods. They largely don't, which is what points the explanation toward the body's response to weight loss rather than toward the specific approach used to produce it.
What researchers mean by successful
This is where a lot of confusion enters, because the research definition is narrower than the everyday one. Wing and Hill proposed defining successful maintainers as people who have intentionally lost at least 10% of their body weight and kept it off for at least one year. The 10% threshold wasn't arbitrary: losses of that size produce substantial improvements in risk factors for diabetes and heart disease.
Note what that definition does and doesn't require. It doesn't require reaching a goal weight, keeping all of the loss, or maintaining forever. Someone who lost 20% and settled at 12% below their starting weight counts as a success by this measure, even though they might describe themselves as having failed. The gap between the research definition and the personal one is wide enough that plenty of people in the successful 20% don't know they're in it.
Losing 20% and settling at 12% counts as success in the research. It rarely feels like it.
What the 20% actually do
The National Weight Control Registry, which tracks people who have maintained substantial weight loss long-term, produces the most detailed picture available of the successful minority. Wing and Phelan's summary of the registry's findings identifies a short list of shared behaviours: a diet low in fat and high in carbohydrate, regular self-monitoring of both weight and food intake, and high levels of daily physical activity. Almost all registry members ate breakfast, and almost all weighed themselves regularly, most commonly once a week.
Two caveats before anyone builds a plan on that list. First, registry members self-select, so this describes what successful maintainers do rather than proving those behaviours caused the success. Second, the dietary composition finding reflects the era the registry was built in and is the least transferable item on the list.
What survives both caveats is the monitoring. Regular self-weighing and regular self-monitoring show up repeatedly across the maintenance literature, not just in this registry, and they have an obvious mechanism behind them: you cannot correct a drift you can't see. How often to weigh yourself covers the frequency question and the psychological cost that comes with it.
Why these numbers are older than they look
Both headline figures come from a body of evidence assembled largely before GLP-1 medications became common, and mostly from people who sought treatment, which is not the same as everyone who has ever lost weight. Treatment-seeking samples skew toward people who have already struggled, so the true population base rate is probably somewhat better than 80%.
It's also worth naming the figure these numbers frequently get confused with. The claim that 95% of diets fail traces to a small clinical study from 1959 involving roughly 100 people, and it has been repeated for six decades without the sample ever getting any larger. The 80% figure is far better supported. Neither is good news, but only one of them is actually evidence.
What this means for you
Base rates describe populations, and you are not a population. Their practical value is in setting the right expectation about what the difficult part is: the losing phase is not where most people come unstuck, and treating maintenance as the easy part that follows the hard part gets the difficulty backwards. The maintenance phase is a different problem with different requirements, not a victory lap.
The second implication is that whatever you do in maintenance has to be something you would still be doing in three years, because three to five years is the window these figures describe. Anything that requires ongoing effort tends not to clear that bar.
How OffRamp helps
OffRamp was built around the one finding from this literature that both replicates and has an obvious mechanism: regular self-weighing. The Regain Radar turns those weigh-ins into a trend line rather than a stream of daily numbers, so a real drift becomes visible while it is still small enough to be worth a small correction. The daily protein floor is deliberately a floor and not a ceiling, because a minimum is something you can still be doing in three years and a restriction generally isn't. There's no calorie target anywhere in the app, and that's a design decision rather than an omission. Anything involving your treatment itself stays a question for your prescriber.
