01
The diet you stick with is the diet that works
Macronutrient composition is largely irrelevant. Adherence is what produces long-term results.
Four of the largest randomized trials ever run — POUNDS LOST (811 adults, 2 years), DIETFITS (609 adults, 12 months), Dansinger's JAMA trial, and Johnston's network meta-analysis — converge on the same finding. Low-fat against low-carb, Atkins against Ornish, Zone against Weight Watchers: macro composition is roughly irrelevant. What predicted success was attendance and self-monitoring at six months. DIETFITS even tested whether genotype or insulin response would help match diets to individuals — neither did. The takeaway is uncomfortable for the diet industry: it's not which calories, it's whether you stay in the game. B.a.D. Coach is built around that one finding.
“Reduced-calorie diets result in clinically meaningful weight loss regardless of which macronutrients they emphasize.”
— Sacks FM et al., POUNDS LOST trial, NEJM 2009
02
Lighter weeks help, not hurt
A deficit broken up by maintenance periods outperforms a continuous deficit.
In a 16-week trial, researchers split 36 men with obesity into two groups. One stayed in a steady calorie deficit. The other ran the same total deficit but in 2-week-on / 2-week-off blocks at maintenance. The intermittent group lost 14.1 kg total versus 9.1 kg in the continuous group — and 12.3 vs 8.0 kg of fat specifically. Their resting metabolic rate also dropped less. The body, it turns out, doesn't punish you for occasionally eating at maintenance. It rewards the break. That's the physiological logic behind a weekly window: a heavier weekend is not sabotage.
“Intermittent group: 14.1 kg total weight loss vs continuous: 9.1 kg (P<0.001).”
— Byrne NM et al., MATADOR study, Int J Obesity 2018
03
Logging is the single biggest lever
Self-monitoring predicts weight-loss success more reliably than any specific diet type.
A systematic review of 22 studies from Lora Burke's group at Pittsburgh confirmed it. Across study types, populations, and time frames, the people who track what they eat lose weight, and the people who stop tracking gain it back. The contents of the diet move the needle a little. The act of paying attention moves it a lot. That's why we obsess over making logging fast — every second of friction is a vote against the one thing the literature is sure about.
“A significant association between self-monitoring and weight loss was consistently found.”
— Burke LE et al., J Am Diet Assoc 2011
04
The body responds to weeks, not days
Fat mass moves over months. The daily scale number is mostly noise.
Kevin Hall's NIH energy-balance models — calibrated against decades of controlled feeding studies — show that a single high or low day produces only a tiny, transient shift in body composition. Adipose tissue turnover is slow; daily scale fluctuations are mostly water and glycogen. The signal lives in the multi-week mean, not the 24-hour total. A daily calorie limit treats every day as if it stands alone — the math says it doesn't.
“The bodyweight response to a change of energy intake is slow, with half times of about 1 year.”
— Hall KD et al., Lancet 2011
05
Punishing yourself is biologically counterproductive
The act of restrictive dieting itself raises stress hormones — independent of any actual weight loss.
A 2×2 experiment with 121 women showed that restrictive dieting raises cortisol output and perceived stress — even when the person isn't losing weight. Chronically elevated cortisol is associated with abdominal fat retention. So an app that hits you with red screens and shame messaging when you go over a daily target may be biochemically working against the goal you bought it for. The weekly window, the no-streak design, the absence of guilt copy — those aren't soft features. They line up with how the endocrine system actually responds.
“Restricting calories increased the total output of cortisol, and monitoring calories increased perceived stress.”
— Tomiyama AJ et al., Psychosom Med 2010
What we don't claim
The literature is solid in places and thinner in others, and we don't want to oversell. There is no head-to-head RCT directly comparing a "weekly target app" against a "daily target app" — that thesis rests on the energy-balance models and the MATADOR trial, not on a direct outcome study. The MATADOR trial itself is one trial, men only, n=36 in the per-protocol analysis. The cortisol-stress link is mechanistic, not an outcome trial — we know dieting raises cortisol; we don't have RCTs proving that non-punishing apps produce more fat loss than punishing ones. And while the self-monitoring evidence is rock-solid, AI-specific coaching trials are early-stage. Where the evidence is strong, we lead with it. Where it's emerging, we say so.
Ready to put it into practice?
The science is clear. Now it's your turn.
Start for free — 10 days, no card needed