Radical Health · Radical Fat Loss · The evidence
The science — the honest version

Backed by research. Honest about its limits.

Every pillar of Radical Fat Loss maps to real, graded evidence — graded the way researchers grade it, and kept only if it survived a deliberate attempt to refute it. Below is each claim with its grade and sources, what the evidence doesn't say, and the full reference list.

01

Calorie deficit & protein

Strong
A calorie deficit drives fat loss; higher protein protects muscle. Adults over 50 retain more lean mass and lose more fat during energy restriction on higher- versus normal-protein diets; the same holds in adults with overweight or obesity.[2, 3]
Moderate
About 1.6 g/kg a day with a little resistance training preserves lean mass — it doesn't build it in a deficit. The protein targets sit inside the evidence-supported range (the ISSN cites ~2.3–3.1 g/kg for trained lifters cutting); a higher-protein deficit plus intense exercise preserved — and even increased — lean mass and strength in a randomized trial.[4, 5, 17]
02

Steps & NEAT

Strong
Everyday movement is the engine of the deficit. Non-exercise activity thermogenesis (NEAT) is the largest and most variable part of daily energy expenditure — it can differ by up to ~2,000 kcal/day between two similar-sized people. That's why we lead with a step floor, not punishing cardio.[1]
Strong · observational
More steps track with longer life. Each additional ~1,000 daily steps is associated with about 15% lower all-cause mortality, with benefit plateauing around 6,000–8,000 steps for over-60s and 8,000–10,000 for younger adults.[26, 27]
03

Resistance training

Strong
Lifting preserves the muscle you'd otherwise lose. Resistance training prevents roughly 93.5% of the lean mass lost during calorie restriction, even at low-to-moderate volume; deficits beyond ~500 kcal/day start to impair retention. It preserves muscle in a deficit — it doesn't build it.[21, 22, 38]
Strong
It matters more with age. Resistance training is the robust, evidence-based intervention for sarcopenia in older adults (where drug therapies have been inconsistent); in one 12-week trial, training plus nutrition cut sarcopenia prevalence from ~35% to 0%. The app's library is deliberately gentle — no training to failure.[6, 7]
04

Fasting & meal timing

Moderate
A fasting window is structure, not a fat-burn trick. Time-restricted eating produces real fat loss versus non-TRE controls (about −1.40 kg fat mass) — but not significantly more than an equivalent calorie deficit. Window length is not the active variable, so we never gamify ever-shorter windows.[10, 11, 14]
Strong
Eating the same calories late costs you. Shifting identical calories ~4 hours later raises hunger and the ghrelin-to-leptin ratio and lowers daytime energy expenditure — a circadian-misalignment effect shown in a controlled crossover trial. That's why we nudge the eating window earlier, framed as appetite quality, never bonus deficit.[12]
Moderate
Earlier eating shows a small edge — at a matched deficit. Early time-restricted eating improved fat-mass and fasting-glucose outcomes at the same calorie deficit, though total weight loss didn't differ (secondary outcomes).[13]
05

Glycaemic levers

Strong · acute
Food order helps, modestly. Eating fibrous veg and protein ~10 minutes before carbohydrates can blunt the glucose peak by up to 44% (iAUC down to −55%), with average glucose unchanged. Real and worth doing — not a weight-loss guarantee.[15, 16]
Moderate
A post-meal walk smooths blood sugar. Better than sitting — and better than walking before the meal (pre-meal walking shows no effect). The window is the first ~0–29 minutes after eating, and 3×10 minutes beats one 30-minute walk.[20]
Strong
One glucose reading can't judge a food. Single-meal CGM responses are unreliable within a person (intraclass correlation 0.17–0.28; most of the variance is within-subject noise). So we aggregate repeated exposures and show a confidence meter rather than a snap verdict.[18]
06

Sleep & behaviour

Strong
Sleep is the quiet lever. In habitual short sleepers, gaining ~1.2 hours of sleep cut energy intake by about 270 kcal/day on its own (measured by doubly-labelled water) — the one lever that subtracts calories without willpower.[28]
Moderate
Logging more, more consistently, tracks with losing more. Self-monitoring shows a dose-response with weight loss — but adherence decays over time, which is exactly why the app is built around following through (pause days, forgiving streaks).[29, 30, 31]
Strong
"If-then" plans work. Implementation intentions — pre-deciding a response to a specific cue — improve goal attainment with a medium-large effect (d≈0.65). The app pre-arms them at your own relapse hotspots.[32, 50]
07

Diet breaks

Moderate
Planned breaks help adherence — not a metabolic "reset." Intermittent energy restriction (two weeks on, two off) improved weight- and fat-loss efficiency in the MATADOR trial. But the most consistent benefit across the literature is adherence: the ISSN finds no body-composition advantage over daily restriction. We frame breaks honestly.[8, 9]
08

The GLP-1 era

Strong
On a weight-loss injection, protect the muscle — don't fear the drug. About 25–38% of GLP-1 weight loss is lean mass, but the fat-to-lean ratio matches diet alone and muscle quality improves. The actionable concern is absolute lean loss without protein and lifting — so we defend protein and strength, and we never call these drugs "muscle-wasting."[33, 35, 37, 38]
Moderate
Appetite suppression and regain need watching. Appetite suppression cuts intake of all macros — including protein, so protein defence matters more, not less — and regain after stopping is asymmetric ("collateral fattening": fat returns faster and more completely than lean). Validated food-noise instruments track the trend. More on using it on a GLP-1 →[41, 45, 49]
What we'll never claim
  • ✕ That a shorter fasting window burns more fat. Window length isn't the variable.
  • ✕ That early eating beats a plain calorie deficit for weight loss. It doesn't pairwise — timing is never a deficit multiplier.
  • ✕ That high protein builds muscle in a deficit. It preserves it.
  • ✕ That a weight-loss injection "wastes muscle." It doesn't — we just protect the muscle you'd otherwise lose.
  • ✕ That "afterburn," fasted cardio, or exercise "snacks" melt fat. They wash out at matched calories — good for fitness, not a fat-loss lever.
  • ✕ That one glucose reading can judge a food. We wait for a pattern.
  • ✕ That AI photo macros are lab-grade. They're fast, honest estimates — never ground truth.
  • ✕ That "7,700 kcal ≈ 1 kg" is exact. A planning estimate, continuously checked against your scale.
Honest caveats

Where the math is
an estimate.

7,700 kcal ≈ 1 kg

The rule behind the deficit bank is a useful planning approximation, not a precise law — real loss varies with water, glycogen and adaptation. The scale-vs-deficit reconciliation exists to catch exactly that drift.

AI photo macros

Photo estimates speed up logging and are meaningfully better than unaided guessing, but they aren't weighed-and-measured intake. Treat them as good estimates — and edit any value before logging.

A single CGM reading

One post-meal glucose curve is too noisy to judge a food. Personalisation has to aggregate repeated exposures — which is why a single reading is never turned into a verdict.

Every claim above is graded the way researchers grade it, and kept only if it survived a deliberate attempt to refute it. Here's the honest part: the evidence backs the method — deficit, protein, lifting, steps, sleep, if-then plans — not a study of this app. There is no prospective outcome trial of Radical Fat Loss itself, and the underlying target math is calibrated to one reference user and generalised by formula rather than validated in a cohort. We'd rather say that than pretend otherwise.

References

The reading list.

The literature behind the method. Links go to PubMed Central, PubMed, or the trial registry where a stable identifier exists; the rest are listed by journal and year.

  1. Levine JA. Nonexercise activity thermogenesis. J Intern Med 2007;262:273–287.
  2. Kim JE et al. Dietary protein & body composition after weight loss in older adults (meta-analysis). Nutr Rev 2016. PMC4892287
  3. Enhanced protein intake & muscle mass in overweight/obesity (meta-analysis). Clin Nutr ESPEN 2024.
  4. Jäger R et al. ISSN Position Stand: protein and exercise. J Int Soc Sports Nutr 2017. PMC5477153
  5. Longland TM et al. Higher vs lower protein during an energy deficit + intense exercise (RCT). Am J Clin Nutr 2016.
  6. Resistance-training prescriptions in older adults with sarcopenia (systematic review/meta-analysis). Aging Clin Exp Res 2025.
  7. Strength training + nutrition vs sarcopenia (12-week RCT). 2025. PMC12295157
  8. Byrne NM et al. Intermittent energy restriction (MATADOR). Int J Obes 2017. PMID 28925405
  9. Aragon AA et al. ISSN Position Stand: diets and body composition. J Int Soc Sports Nutr 2017. PMC5470183
  10. Fasting vs continuous calorie restriction (systematic review/meta-analysis of RCTs). Nutrients 2024;16(20):3533.
  11. Intermittent fasting vs continuous calorie restriction (systematic review/meta-analysis). Nutrients 2022. PMC9099935
  12. Vujović N et al. Late isocaloric eating increases hunger, decreases energy expenditure. Cell Metabolism 2022.
  13. Habe et al. Early vs late TRE at a matched deficit (RCT). 2025. PMC12309031
  14. Time-restricted eating meta-analysis (fat mass −1.40 kg). Nutrients 2024;16(19):3390.
  15. Carbohydrates-last food order improves time-in-range. Diabetes Care 2025;48(2):e15.
  16. Food-order systematic review (PRISMA). Clin Nutr Res 2026.
  17. Protein + resistance training preserves lean in a deficit (RCT). Int J Sport Nutr Exerc Metab 2025. PMID 40796095
  18. CGM within-person imprecision (controlled-feeding RCT, 1,189 duplicate meals). Am J Clin Nutr 2024.
  19. App-based personalised nutrition (ZOE) RCT. Nature Medicine 2024.
  20. Post-meal walking & postprandial glucose (meta-analysis, 8 RCTs). Sports Medicine 2023. PMC10036272
  21. Roth et al. Resistance-training volume & lean preservation. Scand J Med Sci Sports 2023.
  22. Murphy & Koehler. Energy deficit impairs resistance-training gains. Scand J Med Sci Sports 2022.
  23. Exercise modalities for fat loss (network meta-analysis). Frontiers in Nutrition 2025.
  24. Exercise "snacks" (RCT). 2024. PMID 38569204
  25. Exercise-snacks meta-analysis. Frontiers in Cardiovascular Medicine 2025.
  26. Paluch AE et al. Steps & all-cause mortality. Lancet Public Health 2022. PMC9289978
  27. Steps umbrella review. Lancet Public Health 2025.
  28. Tasali E et al. Sleep extension reduces energy intake. JAMA Internal Medicine 2022. PMC8822469
  29. Log Often, Lose More (self-monitoring frequency). PMC6647027
  30. Self-monitoring & weight loss (systematic review). PMC8928602
  31. ≥2 logged eating occasions/day & adherence. PMC6856872
  32. Gollwitzer PM, Sheeran P. Implementation intentions & goal attainment (meta-analysis). 2006 (d≈0.65).
  33. SURMOUNT-1 DXA body-composition substudy. Diabetes Obes Metab 2025.
  34. 22-RCT body-composition network meta-analysis (GLP-1). 2024.
  35. Multi-society advisory on muscle during weight loss. 2025. PMC12125019
  36. Neeland et al. Body-composition heterogeneity review. Diabetes Obes Metab.
  37. SURPASS-3 MRI muscle substudy. Lancet Diabetes & Endocrinology 2025.
  38. Sardeli AV et al. Resistance training preserves lean during calorie restriction (meta-analysis). Nutrients 2018. PMC5946208
  39. LEAN-PREP RCT (protein during GLP-1). NCT06885736
  40. S-LITE: exercise + GLP-1 combination. NEJM 2021.
  41. FNQ food-noise questionnaire validation. Obesity 2025.
  42. Food-noise review. Nature Nutrition & Diabetes 2025.
  43. RAID-FN instrument. Appetite 2025.
  44. Tirzepatide vs semaglutide (network meta-analysis). J Diabetes 2026.
  45. STEP-1 extension (regain after discontinuation). 2022. PMC9542252
  46. Dulloo AG. Collateral fattening / catch-up fat. PMC5945583
  47. Catch-up fat / sarcopenic-obesity review. Rev Endocr Metab Disord 2025.
  48. Postmenopausal weight loss–regain body composition (DXA cohort). Am J Clin Nutr. PMC3155932
  49. Tirzepatide reduces intake of all macros (incl. protein). Nature Medicine 2025.
  50. Marlatt GA, Gordon JR. Relapse Prevention: Maintenance Strategies in the Treatment of Addictive Behaviors. Guilford Press, 1985.

Grades follow an adversarially-verified review: a claim is "confirmed" only if it survived a deliberate attempt to refute it. Reference details are reproduced as published; where no stable identifier was available we list the citation in plain text rather than guess a link.

Honest math.
Show up. Log.

Radical Fat Loss is pre-launch. Join the early list to be first in — one email at launch, nothing before.

Join the early list