Diet · 18 patterns
Every diet, and what happened when somebody held the calories equal
Plate 35
Diet record
Protein, GLP-1, patterns
18
Patterns
human
Results in people first
The patterns, and what the studies found
Ketogenic
Carbohydrate-restricted
When calories were held equal, the edge is real but tiny, and in the best pooled review it points the other way. Nothing here backs the claim that ketosis itself makes you lose much more fat on the same calories.
OpenCarnivore
Elimination
None have been run. Not one controlled feeding study, not one hospital-unit study, not one study with a comparison group, not one trial of any kind. We say this flatly because it is the most important fact about this diet.
OpenVegan and vegetarian
Plant-forward
Almost never, and that is the main catch in this research. The best trials provided food but did not match calories, so part of every result comes from weight loss rather than from the diet itself.
OpenMediterranean
Whole-food
Calories were rarely controlled, and that decides what a study means. PREDIMED did not cut calories and got heart benefits without weight loss. DIRECT did cut calories and got weight loss. Mixing up those two is the most common mistake made with this diet.
OpenPaleo
Elimination
The core claim fails its cleanest test. When weight and the mix of carbs, fat and protein were held steady, cutting grains, dairy and beans made no difference to blood sugar control. Where controlled studies did find benefits, either the calorie control failed and people lost weight anyway, or there was no comparison group.
OpenWhole food versus ultra-processed
Whole-food
Calories were deliberately not controlled, because how much people ate was the thing being measured. What was matched was the food people were offered. The key finding is that the matching fell short in exactly the way that matters.
OpenDASH
Whole-food
This is the rare diet whose main evidence comes from feeding studies with calories controlled. Food was provided, salt was fixed, and body weight was held steady on purpose, in the first study and in every major one after it. So the blood pressure and cholesterol results are not weight loss results. The top blood pressure number fell 5.5 mmHg against a typical American diet, and 11.4 mmHg in people who had high blood pressure.
OpenLow fat
Macronutrient-defined
At the same calories, cutting fat leads to more body fat loss than cutting carbs. That is the opposite of the popular claim, and the effect is tiny either way. Against other real diets of similar effort, there is no weight advantage at all. The diet's real calorie finding came from people eating freely: in a hospital unit, people offered a low fat plant-based diet ate 689 calories a day less than on a keto diet.
OpenAtkins and non-ketogenic low carbohydrate
Carbohydrate-restricted
With calories set the same, the difference against a balanced carb diet is about one kilogram at three to eight months, and it shrinks toward nothing by two years. The difference is real, and it is a kilogram. Where food was actually provided, rather than just prescribed, the metabolic edge runs the other way, by tens of calories a day.
OpenMIND
Whole-food
No calorie-controlled feeding study of this diet exists, and none is likely, because memory results take years and feeding studies run for weeks. What does exist is a three-year trial in which both groups got a mild calorie cut and weight loss support. So calories were matched between groups by design, and the trial found no difference in memory and thinking or in brain scans.
OpenLow FODMAP
Elimination
Calories are not what matters here, and they were not the question. What matters is whether the comparison diet was as demanding, as believable and as well supported as the real one, and the answer changes the result. Against a typical diet, with all food provided, the effect is large. Against a fake diet that cut a similar number of foods, the main result just missed being clear. Against standard diet advice, the diet wins on bloating and ties on everything else.
OpenGluten free without coeliac disease
Elimination
Calories are not the question, and no calorie-controlled study exists or is needed. The question is whether gluten is the cause, and that has been tested the right way: people are given gluten or a matching dummy without knowing which. The answer has shifted over fifteen years, from yes in the first study to a small minority in the pooled review. In the study that tested both, fructans beat gluten as the trigger.
OpenHigh protein
Macronutrient-defined
Protein does not change how much fat you store at a given calorie intake. It changes what the rest of the weight is made of, and it raises the energy you burn at rest by a couple of hundred calories a day at the top end. Both are real, and neither is a loophole.
OpenCalorie counting and flexible dieting
Macronutrient-defined
Constantly, and that is the point. When intake is truly fixed, the result is predictable, and it is the foundation the rest of this site is built on. For anyone doing this without supervision, the finding that matters is the size of two gaps: between the cut you plan and the cut you achieve, and between what you eat and what you report. Both have been measured.
OpenCommercial weight loss programmes
Macronutrient-defined
Never, and here that is the right design, not a weakness. The program is about changing behaviour, so what matters is what people actually eat when nobody is watching. Controlling their food would remove the very thing being tested.
OpenBlue Zones
Plant-forward
Never, and it cannot be. There is nothing here to control. No trial of any Blue Zone diet has ever been run, the diet data looks backward, and the most serious criticism of this field holds that some of the age records are wrong.
OpenAlkaline diet
Elimination
Calories are not the disputed factor here. Acid load is, and it has been controlled again and again in studies that measure every bit of calcium in and out, with calcium, phosphate, sodium, protein and calories matched between groups. The first half of the idea holds up: raise the acid load and urine calcium rises. The second half fails: the body's total calcium balance does not change.
OpenBlood type diet
Elimination
No calorie-controlled study of this diet exists, and neither does any trial built to test it. The only trial evidence is a side analysis inside a study run for another purpose, and it found blood group made no difference to any result it measured.
Open18 patterns, 188 sources, every one checked against the original before it was published. A study that matched calories is measuring the diet. A study where both groups ate freely is measuring the diet and your appetite together. Both are real questions with different answers. Blending them is how the same research gets used to prove opposite things. Compare them side by side.
Grouped by what they actually restrict
Most diets that look like rivals are the same restriction wearing different branding, and most arguments between them are arguments between two members of one family. Grouping them this way makes that visible.
Carbohydrate-restricted
- Ketogenic13 in people
- Atkins and non-ketogenic low carbohydrate13 in people
Elimination
- Carnivore1 in people
- Paleo11 in people
- Low FODMAP9 in people
- Gluten free without coeliac disease8 in people
- Alkaline diet4 in people
- Blood type diet6 in people
Macronutrient-defined
- Low fat12 in people
- High protein8 in people
- Calorie counting and flexible dieting11 in people
- Commercial weight loss programmes8 in people
Plant-forward
- Vegan and vegetarian10 in people
- Blue Zones7 in people
Whole-food
- Mediterranean4 in people
- Whole food versus ultra-processed2 in people
- DASH12 in people
- MIND9 in people
Before you pick one
The studies that matched calories keep finding smaller differences between these patterns than the people arguing about them expect. The one thing that reliably changes what the weight you lose is made of is protein. It is also the one most often got wrong, in the direction that costs you muscle. That is a position rather than a finding, so it is labelled as one.
The Archive’s position
This is opinion, not a finding. Everything else on this page is the evidence.
Eat one gram of protein per pound of your goal bodyweight, every day, and treat it as the floor, not the target. If you want to weigh 200 pounds, you eat 200 grams, whether you weigh 175 or 260 right now.
Why we hold it
- The RDA, the official daily recommendation, is a floor against running short, not the best amount. It was worked out by counting protein in and out in inactive adults, and it answers one question: when do you stop losing protein from your body? Almost nobody reading this is trying to just not waste away.
- The authors of the paper most often used to argue against this recommended the number themselves. The 2018 review that found the 1.62 g/kg plateau says in its own discussion that, since its likely range reached 2.20, it may be wise to recommend about 2.2 g/kg a day for people trying to get the most muscle from training. That is one gram per pound, from the very source usually quoted against it.
- A second, separate method lands on the same top figure. Tracking a tagged protein building block in the breath, instead of counting protein in and out, put the top of the likely range at 2.2 g/kg a day in trained lifters. Two methods with nothing in common reaching the same ceiling is worth more than either one alone.
- Nearly everyone who cares about this is eating less to lose weight, and that changes the answer. In the two controlled studies that tested it directly, roughly 2.3 to 2.4 g/kg beat 1.0 to 1.2 g/kg on lean mass by a wide margin. In one of them, the higher protein group gained muscle while losing fat.
- Goal weight is a workable stand-in for lean mass. Sports nutrition research scales protein to lean mass, which nobody knows without a scan. Goal weight gets close to it, and it stops a person at 300 pounds from working out a target their muscle does not need.
- Being wrong costs more one way than the other. In healthy people with normal kidneys, four months at 2.6 to 3.3 g/kg changed no kidney, liver or cholesterol marker. Eating too little protein while cutting calories reliably costs lean tissue you then have to rebuild.
- It is a rule a person can actually follow. A number you can work out in your head and hit without a spreadsheet gets followed. A more precise number that needs a body scan does not.
Where this goes beyond the evidence
The central estimate is 1.6 g/kg, not 2.2, and no pooled review has ever found a plateau at 2.2. The 2.2 figure is the top of the likely range on a result that was not clear, which is a reason to doubt the ceiling, not proof of a higher one. The sharper problem is a study that tested this directly. 39 adults eating 40 percent less were given 0.8, 1.6 or 2.4 g/kg for three weeks. Both higher groups beat the RDA at protecting lean mass, but 1.6 and 2.4 could not be told apart. That is the closest thing to a head-to-head test of this rule, and it found no advantage to the higher number. The honest summary is that the gap between 1.6 and one gram per pound is roughly 40 to 60 grams a day, at an intake where no harm has been measured in healthy kidneys. Choosing the top of the range buys insurance against real uncertainty. It does not buy a measured extra effect, because in the one study built to find one, there was none to find.
Who this is not for
Anyone with reduced kidney function or chronic kidney disease. The review that cleared high protein for kidney function only included adults without kidney disease, and its authors limit the conclusion to them, so it does not carry over. The rule also breaks at the edges of its own math. Someone at 150 pounds aiming for 200 would be eating close to 3 g/kg of what they weigh now, past every measured plateau. Someone very heavy aiming far lower lands below the range the research on eating less recommends. It ignores height entirely. And 200 grams is 800 calories, which squeezes a smaller person's calorie cut hard. It is a good default, not a law of physics.
Also in this section
How much protein
The official daily amount (the RDA) and why its own authors called it a floor for staying even. The 1.6 g/kg level where gains flatten out, and the two caveats that get dropped when it is quoted. And the studies of people eating less than they burn, where the answer changes. Six measured intakes, each with the group of people it came from.
OpenGLP-1s, muscle and hair
Cutting appetite works, and then people stop eating rather than eat differently. What the body scans measured, why the share of weight lost from muscle and other lean tissue is the same on a dummy shot, and where the hair loss reports actually sit.
OpenThe patterns compared side by side
All 18 patterns in one view. The studies that matched calories are shown apart from the ones where people ate freely. That one difference explains most of the seeming disagreement in nutrition, and almost nobody separates them.
OpenHow this section is written
Every diet result here comes with one thing attached: whether the study controlled how much people ate. A study that matched calories between groups is measuring what the diet is made of. A study where both groups ate freely is measuring the diet and its effect on appetite and on sticking with it, together. Both are real questions and they have different answers. Blending them is how the same research gets used to prove opposite things.
Every pattern page also names who the diet is wrong for. It also reports the test results that move the wrong way, next to the ones fans lead with. The site is built so a pattern page cannot be published without them.
Where the owner of this site holds a position, it appears in a block like the one above, labelled as opinion. It includes the part where it goes past what has been measured. That part is required. A position that cannot name its own weak point is marketing, and this site is not that.