The Supplement Sort: What Eight Research Teams Found When They Tested Everything on the Shelf

You have a shelf full of bottles. Maybe a cabinet. A monthly auto-ship for at least two of them.

Somewhere between the creatine and the fat burner sits a tub of BCAAs someone at the gym told you was essential. Next to it, a test booster that promised visible results in thirty days. The total runs somewhere between a hundred and two hundred dollars a month. And if you are being honest, you are not sure which of them is doing anything.

You are not alone. It is the most common question in the supplements space, and the most poorly answered. Tier lists from creators who sell the products they rank. Affiliate articles that review brands without asking whether the category works at all.

So we collected it. Eight independent research teams pooling every qualifying trial for their supplement category. Over 12,435 participants across deduplicated studies. Every supplement category the mass fitness audience buys — tested, ranked, and verified through a process you can inspect on the Skeptic Protocol page. What we found is shorter and cheaper than you expect.

The supplement with the strongest evidence of all is creatine. Not close. Not debatable.

143 controlled trials. Zero disagreement between studies. When researchers pooled the data — over three thousand people — they found a consistent gain of about 0.8 kg of fat-free mass.

And the tools that measured it did not stop at scales. A separate research team used MRI and CT scans across ten studies. The result: a 96% probability that the muscle tissue itself was genuinely larger. An MRI does not care about water. If the tissue is bigger on the scan, the tissue is bigger.

A third line of evidence closes the case. Across 53 studies, creatine improved upper body strength. Water retention does not make you stronger. Three methods of proof. One answer. The muscle is real.

Now the part that rewrites who creatine is for. Under forty: plus 0.89 kg. Over forty: plus 0.87 kg. The gap between groups was so small it was statistically meaningless.

Adults aged 57 to 70 gained even more. Creatine is not a young person's supplement. It works at every age tested.

And the cost: about $0.10 a day. Plain creatine monohydrate — the form used in 89 of 95 studies. The alternatives cost more than double and had three studies between them.

The best supplement on the shelf is the boring one nobody markets. If the best supplement costs ten cents a day, what does the rest of the shelf look like under the same evidence standard?

Do fat burners work?

No. And the data is not ambiguous.

When researchers pooled 21 controlled studies and over two thousand people, the result was the same on every measure — total weight, fat mass, muscle retention. All of them crossed zero. None could be distinguished from doing nothing.

Taking them longer did not help. Eight weeks. Twelve weeks. Twenty weeks. At every time window, zero.

The one promise on every label — boosting your metabolism — showed an effect of 0.018. Not rounding to zero. Already there.

Five of seven groups saw their metabolic rate going down, not up. The product that promises to speed up your engine may be slowing it.

And then the comparison nobody in the supplement space talks about. Diet and exercise were tested head-to-head against fat burner supplements. The free option won. The gap was too large to be coincidence.

For muscle retention, the difference was even wider. You have been paying forty to sixty dollars a month for a product that loses to something you can do for nothing. And 43% of participants experienced side effects — heart rate spikes, blood pressure changes, sleep problems — for zero proven benefit.

The sweat, the heart rate, the focus — those sensations are real. They are side effects of stimulants, not evidence of fat leaving your body.

Here is where the story turns. The only active ingredient in most fat burners that has any evidence at all is caffeine. And caffeine does have a real effect — just not on fat.

Four independent research teams measured a small but confirmed strength boost across thousands of participants. It blocks the tiredness signal in your brain, letting your nervous system recruit more muscle fibers per rep. That effect is real and replicated.

But when the same molecule is repackaged as a thermogenic fat burner, the body composition effect is zero. Same caffeine. Two labels. One promise kept, one empty.

The reliable benefit — slightly stronger lifts — is unglamorous and rarely on the label. Your fifty-dollar fat burner is repackaged caffeine. A cup of coffee costs less than a dollar.

If caffeine is the only ingredient that does anything in your fat burner — and most protein powders are no better than a sugar pill — what actually works in the protein aisle?

Start with the failure rate. Thirteen types of protein powder. Whey, casein, soy, pea, collagen, and eight more. All tested simultaneously across 78 studies and nearly five thousand people.

11 of 13 types showed zero benefit over a sugar pill. Not a small benefit. Not a borderline result. Nothing.

The gap most people try to fill with protein powder is about 31 grams — the difference between what most people eat and the evidence-backed threshold of 1.6 g/kg. That is one extra portion of food.

The gym rule of one gram per pound overshoots the evidence by about forty percent. Protein powder works because most people fall short. If food covers it, the powder has nothing left to add.

Now the paradox that breaks the framework.

BCAAs deliver three of nine essential amino acids. They flip the switch that starts muscle building. But without the other six, your body has to cannibalize existing muscle for the missing parts. More than twenty studies. The supplement meant to build muscle tears it down for parts.

Collagen has 2% leucine. It is missing tryptophan entirely. By every protein quality scoring system the supplement industry uses, collagen should be the worst performer on the shelf.

Across 19 studies and nearly eight hundred people, collagen produced real gains in lean mass and strength. Every study measuring strength pointed the same direction. Zero disagreements.

The protein that every scoring system says should fail outperformed every premium powder.

The resolution: completely different biological pathways. BCAAs attempt muscle protein synthesis and fail because six amino acids are missing.

Collagen does not use that pathway at all. Its amino acids — glycine, proline, hydroxyproline — build the scaffolding around your muscles, not the muscle fibers themselves. Whey builds the walls. Collagen builds the frame that holds the walls up.

The gains are real — but they may be connective tissue rather than muscle fiber. Different job, different pathway, different outcome.

If even the protein quality scoring system is wrong, how do you know which supplement labels to trust?

Here is what we do not know: nobody has tested creatine, protein, and caffeine together as a combined stack. Every study tested one supplement against a placebo. The recommended stack is assembled from individual verdicts, not tested as a unit.

That gap matters. And it is not the only one.

When researchers audited what is actually inside test booster bottles, the findings were worse than the marketing. 75% of ingredients had never been tested. Not mixed results. Not weak data. Zero published evidence for most of what people are swallowing.

One in ten ingredients had data suggesting they might actually lower testosterone. The pill promising to raise your levels could be pushing them the wrong way.

Even the ones that do raise testosterone stay inside your body's normal range. Nobody has shown that a small, normal-range bump builds muscle. Raising T with an herb and growing tissue are two different jobs.

That chain from bottle to bicep has a missing link nobody tested. The supplement that does not claim to change your hormones works. The ones that do claim it don't.

The one ingredient that delivers — ashwagandha — works through stress and cortisol pathways, not through testosterone. It is a recovery supplement in a testosterone booster costume.

Fish oil follows the same pattern. Zero effect on muscle, tested fifteen different ways. Every dose, every duration, every age group. The inflammation from training is part of the growth signal — suppressing it does not help.

The pattern holds across every category. Now: the single framework that makes every supplement decision simple.

All the numbers from the previous sections resolve into one framework. Not an opinion. Not a tier list from someone selling you a stack. What eight independent research teams found when they each took one category and tested every qualifying trial.

WORKS: Creatine (143 trials, zero disagreement, real tissue on imaging). Protein powder (closes the 31-gram gap when food falls short). Caffeine (small, confirmed strength boost — same molecule whether it comes from a pre-workout tub or a jar of instant coffee).

CONDITIONAL: Collagen (19 studies, real gains — but through connective tissue, not muscle fiber growth. A complement, not a replacement).

FAILS: BCAAs (redundant if protein is adequate). Fat burners (every measure crossed zero). Fish oil (zero effect, tested fifteen ways).

UNTESTED: Test boosters (most ingredients never tested. The few that raise T do not build muscle).

3 of 8 categories work. That is the ranking from over 12,435 participants.

Now the price-evidence inversion that should change how you spend your money. The supplement with the strongest evidence — creatine — costs about $0.10 a day. The supplements with the weakest evidence — test boosters and fat burners — cost one to three dollars a day.

The industries with the biggest marketing budgets have the weakest data. The entire pricing structure is backwards.

The evidence-backed stack: creatine, protein when meals fall short, caffeine if you want the edge. $30 to $40 a month.

Whatever you spend beyond that is paying for marketing, not outcomes.

Two honest gaps. Nobody tested the full stack as a unit — creatine plus protein plus caffeine together. Each verdict is strong on its own. But the combo is assumed, not proven.

And 80% or more of the people in these studies were men. The data on women is thin across the board. One exception: creatine was tested in both sexes and worked the same way.