Calvin
How does an adaptive calorie target work?
An adaptive target works backwards from the scale. It compares your logged intake against your smoothed weight trend over about 3 weeks, solves for the maintenance calories that would explain the result, and adjusts. That is why it survives under-logging: the trend is measured, while your log is only an estimate.
Why doesn't a calorie formula stay accurate?
Because a formula predicts an average, and you are one person. Mifflin-St Jeor was built from 498 adults and explains about 71% of the variance in resting metabolic rate, which leaves real room for error. Activity estimates add more. The result drifts from what your own body actually spends.
Two people the same height, weight, age and sex can differ meaningfully in what they burn at rest, and the standard calorie equations have no way to see that. Layer an activity multiplier on top — a single number standing in for how you move all week — and the error compounds.
None of that makes the formula useless. It makes it a starting point that needs correcting against something measured.
Why is the 3,500-calories-per-pound rule wrong?
It assumes your metabolism never responds, and it does. The American Society for Nutrition recommended abandoning the rule in a 2012 consensus statement, because energy expenditure falls as you lose weight, so a fixed deficit yields progressively less loss than the arithmetic predicts. NIH's Body Weight Planner models that decline instead.
The NIH Body Weight Planner, developed at NIDDK, simulates how expenditure and body composition shift in response to a change in intake — which is why its projections curve rather than run in a straight line. The practical consequence: a deficit that worked in month one is a smaller deficit by month three, without you changing anything.
What does an adaptive target need from you?
Three weeks of weigh-ins and a log that is consistently wrong rather than randomly wrong. Consistency matters more than accuracy: if you under-count by 15% every day, the maths absorbs it into your maintenance estimate. Skipping days at random is what actually breaks the calculation.
| Habit | Effect on an adaptive target |
|---|---|
| Weighing most mornings under the same conditions | Essential — this is the measured signal |
| Logging every day, imperfectly but the same way | Fine — a constant bias cancels out |
| Logging carefully on weekdays, not at all at weekends | Breaks it — the bias is not constant |
| Adding back exercise calories from a watch | Counterproductive — re-adds the guesswork |
| Changing the target manually every few days | Prevents the estimate from ever converging |
This is the part people find counterintuitive. An adaptive target does not need you to be an accurate logger. It needs you to be a repeatable one, because it is solving for the offset between what you record and what the scale reveals.
How is this different from what other trackers do?
Most trackers set a target once from a formula and then add back the calories your watch says you burned. An adaptive target ignores that loop and reads the outcome instead. MacroFactor built its whole product on this idea, and Calvin uses the same principle for its free target.
The distinction matters most for people whose logging is honest but incomplete — which is nearly everyone after the first fortnight. A formula-based target treats your log as truth and your body as predictable. An adaptive one treats the weight trend as truth and your log as evidence.
Common questions
Does an adaptive target still work if I log badly?
It works if you log badly in the same way every day. A consistent 15% under-count is absorbed into your maintenance estimate, because the scale reveals the true energy balance regardless. Logging carefully some days and not at all on others is what breaks it.
How long until an adaptive calorie target settles?
Around 3 weeks of weigh-ins and logs. The calculation needs enough data for daily weight noise, which runs 2 to 4 pounds, to cancel out. Before that the estimate exists but moves a lot; after it, the target changes slowly and for real reasons.
Should I eat back the calories my watch says I burned?
Not with an adaptive target. Wearable activity estimates carry large errors, and adding them back re-introduces the guesswork the method exists to remove. An adaptive target already includes your typical activity, because your weight trend reflects everything you actually did.
Why did my calorie target go down?
Usually because your weight trend moved faster than your logged intake predicted, so the estimate of your maintenance calories fell. Energy expenditure also declines as body weight comes down, which is a real physiological effect and the reason fixed targets stop working.
Sources
- NIH Body Weight Planner (NIDDK). Dynamic model of energy expenditure and body composition, developed by Kevin Hall.
- USDA announcement of the NIH Body Weight Planner. Describes why the 3,500-calorie rule does not account for metabolic slowing.
- Hall & Guo, "Obesity Energetics: Body Weight Regulation and the Effects of Diet Composition". Peer-reviewed background on adaptation during weight loss.
Calvin's target adapts, free
After about three weeks of weigh-ins, Calvin back-calculates your maintenance calories from your weight trend. No subscription for it.
Coming soon to the App Store, for iPhone and Apple Watch. See what Calvin does.
This page is general information, not medical advice, and Calvin is not a medical device. Talk to a doctor or a registered dietitian before making significant changes to what you eat, particularly if you have a medical condition.