What is the body’s “set weight,” and why do diets fail to change it permanently?
The Short Answer
Your body defends a particular weight range. One model calls that a set point or set weight; it is one of three, and not the settled one.[1] When you lose weight through calorie restriction, your body fights back: metabolism slows beyond what the smaller body explains, hunger rises as ghrelin increases, and the body becomes measurably more efficient.[2] And it doesn’t let go: across 29 long-term studies, more than half of lost weight was regained within two years and more than 80% by five.[3] The body treats your set weight as correct and weight loss as a problem to fix.
What sets the set point is a genuinely open question, and this companion takes one answer seriously without pretending it’s settled: the carbohydrate-insulin model, which puts chronically elevated insulin at the center.[4]
Something to Sit With
If your body is defending a set point, what would it take to change what it’s defending, rather than just fighting against the defense?
What if the goal isn’t weight loss, but a changed metabolic environment that makes weight loss the natural consequence?
Learn More
The Body’s Defense System.
Your body isn’t passive about weight. In the lipostat model, leptin signals current fat stores to the brain, the hypothalamus compares that signal against a target, and deviations trigger corrective responses.[1] The heritability figures suggest how much biology is involved here: Friedman puts obesity heritability at 0.7-0.8, comparable to height and higher than schizophrenia, breast cancer, or heart disease[1], while Speakman’s review gives 50-70% and cautions that twin studies overestimate it.[5] The measured genetics are thinner than that sounds, and both figures are real: a genome-wide study of 123,865 people of European ancestry found 32 significant variants explaining 1.5% of the variation in BMI.[5]
One caveat before the mechanisms, because it shapes how much weight to put on the phrase “set point.” It is one of three competing models, and it is not the winner. The settling point model holds that weight stabilizes from environmental factors with no active defense of any target; the dual intervention point model proposes upper and lower boundaries where regulation kicks in with passive drift between them. No pure model is adequate, and gene-by-environment interactions dominate.[5] Set point earns its place here because it is the model that best explains the specific thing this companion is about, weight regain after dieting, while settling point better explains the environmental contribution.[5] Read “set point” as the most useful lens for this question, not as a measured quantity in your body.
The defenses it describes are these:
Metabolic adaptation. Once you have lost weight, your body burns less than the smaller body alone explains.[2] Research on Biggest Loser contestants by Fothergill and colleagues (2016) put a number on how long it lasts: a metabolic adaptation of about 499 fewer calories a day, still present six years after the competition.[3] The detail that unsettles people is what happened to those who regained the weight: their resting metabolic rate stayed suppressed anyway, and the adaptation had actually worsened over time.[3]
Hormonal shifts. Weight loss increases ghrelin (hunger hormone) and decreases leptin (satiety hormone).[2] You feel hungrier after losing weight than you did at the heavier weight. This isn’t weakness: it’s biology.
Thermic reduction. Your body becomes more efficient at using energy, not at extracting more of it from food. After a 10% weight reduction, sympathetic nervous system activity falls by around 40% and skeletal muscle does roughly 20% more work per calorie at low intensity, so the same walk costs you less, which sounds like a benefit and isn’t.[2] You also move less without deciding to, as spontaneous activity drops away.[2]
These adaptations made sense evolutionarily: in an environment of scarcity, defending against weight loss meant survival. But the defense is lopsided, and that asymmetry is the part that matters in an environment of abundance: the body resists weight loss far harder than it resists weight gain. Overfeeding does raise energy expenditure and sympathetic tone, but those responses dissipate after a few months, while the adaptations to weight loss persist for years.[3] Adaptive thermogenesis of resting expenditure shows up with weight loss and not with overfeeding or refeeding at all.[3]
The evolutionary account is that the lower boundary was under strong selection from starvation, while the upper boundary once reflected predation risk, a pressure removed roughly two million years ago, letting the ceiling drift upward ever since.[3]
Why Calorie Restriction Fails.
The standard diet advice (eat less, move more) works temporarily and fails long-term. Not because there’s a dial you failed to turn: the reason is the defense already described, which is measurable and which outlasts the diet by years. Resting expenditure drops by more than the smaller body accounts for, hunger rises, movement falls away, and muscle does more work per calorie.[2] Set point is the model that best accounts for that pattern of regain after dieting[5], an explanation of what’s observed, not a quantity sitting in you that the diet left untouched.
The usual picture for this is a thermostat: open a window and the room cools, but the furnace runs until it’s back to 72°. It’s a useful image of the pattern (push the system away from where it sits and it pushes back), and it’s worth being clear that the thermostat is a metaphor rather than a component. There’s no dial. What there is, is a set of responses that switch on when weight falls and stay on for years.
Calorie restriction is the open window. The set point, on this model, is the thermostat.
The regain evidence is worth stating precisely, because the popular versions of it are wrong in a specific way. A meta-analysis of 29 long-term studies found that more than half of lost weight returned within two years, and more than 80% by five.[3] That is a measure of how much of the weight comes back, not the share of people who fail, which is the form the statistic usually takes when quoted. And the failure isn’t a willpower deficiency; it reflects biological, behavioral, and environmental forces, which is also why modest sustained loss still pays: in the Diabetes Prevention Program, a 7% loss cut diabetes incidence by 58% even though half of it was regained within three years.[3]
There is a real version of “you end up worse off,” though it isn’t about the number on the scale. Regain preferentially restores fat over lean tissue, so each weight cycle shifts body composition toward a higher fat percentage even at the same total weight.[3]
What Sets the Set Point?
Here the honest answer is that nobody is certain, and it matters that you know that before reading on.
The account below is the carbohydrate-insulin model. Its physiology is not in dispute: insulin is the dominant hormone controlling fat cell metabolism, elevated insulin suppresses the release of stored fat, and high-glycemic carbohydrates produce a larger insulin response than protein or fat.[4] What remains debated is the causal claim built on top of that physiology: that this is the primary driver of adiposity. The human evidence is inconclusive.[4] What follows is the model as its proponents state it.
Chronically elevated insulin promotes fat storage over fat oxidation, and suppresses the release of fat already stored.[4] The model’s most counterintuitive move is to reverse the usual arrow: it proposes that overeating is a consequence of increasing adiposity rather than its primary cause.[4]
Insulin resistance compounds it. When cells respond less to insulin, the body produces more of it to achieve the same effect, more fat-storage signal for the same meal.[4]
The strongest support is genetic rather than dietary. A bi-directional Mendelian randomization study found that higher genetically determined insulin secretion strongly predicted BMI (β = 0.098, P = 2.2 × 10⁻²¹), while genetically determined BMI did not predict insulin secretion. The exposure is specific: insulin measured 30 minutes after oral glucose, carbohydrate-stimulated rather than fasting.[4] The asymmetry is the argument: one direction alone would be an association.
Two qualifications travel with that result. The review that relays it was written by two of the study’s own co-authors, so a review reporting the study is one team rather than two sources. And a competing analysis reports the mirror image, with BMI predicting fasting insulin and the reverse direction null. The dispute is about which exposure counts rather than about the arithmetic, since carbohydrate-stimulated insulin and fasting insulin are not the same measure; the standings are unequal as well, the first peer-reviewed and the counter-analysis an unpublished preprint.[4] In rodents, high-glycemic diets produced excess adiposity even under calorie restriction, with diet composition affecting body composition independent of intake.[4] Whether that transfers to humans is precisely the open question.
Changing the Set Point.
If the set point is driven by hormonal environment, changing the set point requires changing that environment:
Intermittent fasting. Periods without eating allow insulin to fall to baseline, and the insulin evidence is suggestive rather than settled. Across six randomized trials, intermittent restriction matched continuous restriction for weight loss, and two of them found substantially larger reductions in insulin resistance from the intermittent arm (−24% versus −4%), independent of body fat change; a review of these regimens finds the reductions in insulin resistance comparable overall.[6] An isocaloric crossover trial points the same way, in eight men with prediabetes: early time-restricted feeding improved insulin sensitivity and beta cell responsiveness with no significant weight difference between arms.[6] Whether that lowers a set point is not something the evidence speaks to; what it shows is that the metabolic benefit isn’t merely a by-product of losing weight.
It also differs from continuous restriction in the hormonal direction it pushes. Fasting raises norepinephrine and growth hormone and can increase energy expenditure; chronic restriction does the opposite, dropping sympathetic activity and, in the Minnesota Starvation Experiment, basal metabolic rate by about 40%.[6]
Reducing refined carbohydrates. Refined carbohydrates trigger a larger insulin response than protein or fat, so reducing them reduces insulin exposure.[4] That much is the undisputed physiology; that it lowers a defended weight is the model’s claim rather than a demonstrated result.
Time. Whatever the mechanism, the defense is measured in years rather than weeks: the Biggest Loser adaptation was still present at six years.[3] Any approach that assumes the body will quickly accept a lower weight is arguing with that timescale.
The Patience Required.
If you’ve been at a higher weight for years, your body has defended that weight for years, and the measured persistence of that defense is in years too.[3]
This is why sustainable approaches feel slow. Whether a defended weight can be reset, and whether insulin is the lever that resets it, is the open question this companion has tried not to answer for you. What is not open is the shape of the opposition: it is biological rather than a failure of will,[3] it outlasts the diet by years,[3] and aggressive restriction is the approach that provokes it most directly.[6]
The practical consequence doesn’t depend on which model turns out right. A modest loss you hold onto is worth more than a large one that comes back: in the Diabetes Prevention Program a 7% loss cut diabetes incidence by 58%, and it did so even though half the weight was regained inside three years.[3]
Further reading:
- Fothergill E et al., “Persistent metabolic adaptation 6 years after ‘The Biggest Loser’ competition”[7] — Obesity, 2016
- Ludwig DS & Ebbeling CB, “The Carbohydrate-Insulin Model of Obesity: Beyond ‘Calories In, Calories Out’”[8] — JAMA Internal Medicine, 2018
- Fung J, The Obesity Code[9] (2016) — On set point and insulin
- Sumithran P et al., “Long-term persistence of hormonal adaptations to weight loss”[10] — New England Journal of Medicine, 2011
Written by AI, directed and approved by me. How this is made →