SLThe Science LibraryRecovery science in three layers: research, mechanism, protocol.RCThe Recovery CoachAsk about an injury. Get mechanism first, and a hard stop at diagnosis.PLThe Protocol LabMove the variables. Watch the physiology respond.NUThe Nutrition DeskFuel as chemistry: what to eat, when, and how much for your actual body mass.TEThe Truth EngineSupplements and treatments audited to a verdict. Same four questions, every time.GAThe Global Recovery AtlasThe world's recovery traditions, each on its own terms and then on the evidence.MBThe Mind-Body WingPain is an output of the brain. Breathwork is vagal arithmetic.FNField NotesOne athlete, one protocol, honest numbers — including the null results.
TEModule 04 · The evidence auditor

The recovery industry sells certainty. This is the audit.

Every supplement, device, and modality here gets the identical four questions: what is claimed, what the molecule or the modality actually does, what the literature supports, and a verdict on the record. Same rubric for a hundred-year-old tradition and a device released last month.

Nothing gets a pass for being trendy and nothing gets dismissed for being old. Several of these fail, and the failures are stated in the headline where you can find them.

Audits published

14

Verdicts that are not a yes

7of 14

Scoring axes

4

References cited

74

The method

Four questions, asked in the same order, every time

A rubric is only worth anything if it is applied to the things you like as strictly as to the things you do not. So the order never changes and the axes never change.

01The claim

What is actually being promised?

Quoted in the industry's own voice, from labels, product pages, and studio copy. Not paraphrased, because the paraphrase is always more reasonable than the original.

02The chemistry

What does it do in the body?

Named molecules, named pathways, real timescales, and the physics where physics decides it. A story about fascia has to survive fascia's tensile modulus.

03The evidence

What does the literature support?

Study designs, populations, sample sizes, effect sizes, and where the research is weak: industry funding, surrogate endpoints, untrained subjects, impossible blinding, n of eight.

04The verdict

So is it worth it?

One tier, one headline, on the record, with who it is actually for and who it is not. No hedging into uselessness.

Editorial position

Negative verdicts are the point. A reviewer who never says no is not a reviewer, they are a catalogue.

Five of the audits below fail, and two of those fail on their central mechanism rather than on a technicality. Nothing here gets a pass for being new and nothing gets dismissed for being old: creatine and cupping answer the same four questions and are graded on the same four axes.

Where the literature genuinely disagrees with itself, the audit says so instead of picking the tidier side. Where a verdict rests on one small study, the audit names the sample size in the sentence that cites it.

The comparison matrix

All fourteen, side by side, sortable

Click any column header to sort. Click a verdict tier or a category to filter. The scores are the audit's own, and the composite weighting is printed under the table rather than kept in a drawer.

Categoryshowing 14 of 14
TE-01Creatine monohydrateThe most tested ergogenic aid in sports nutrition, and it costs about a dime a day.Supportedsupplement
95
93
70
97
87
TE-13Beta-alanineSupported, and narrow: if your effort does not last between one and four minutes, this is not your supplement.Supportedsupplement
90
85
45
80
72
TE-14Ice baths after liftingExcellent for playing again tomorrow. Measurably bad for getting bigger.Conditionalmodality
80
75
45
85
69
TE-11NSAIDs for training sorenessThey work on the pain. That is the problem: the pain pathway is also the remodelling pathway.Conditionalpharmacological
90
75
60
35
62
TE-12Foam rollingIt does something real, and it is not what the phrase myofascial release implies.Conditionalmanual therapy
40
60
32
90
57
TE-04Magnesium supplementationFixes a real deficiency, does very little without one, and almost certainly will not stop your cramps.Conditionalsupplement
80
55
28
68
53
TE-10Compression garmentsAbout one point of soreness on a ten-point scale, and very little a coach could measure.Overstateddevice
45
60
25
50
45
TE-08Percussive massage devicesReal range of motion, real pain relief, both largely gone within the hour, and no, it is not breaking up scar tissue.Overstateddevice
45
55
28
45
43
TE-02Collagen peptides + vitamin CProbably useful for tendon, for a reason that has nothing to do with what is written on the tub.Conditionalsupplement
60
45
35
40
42
TE-05Tart cherry juicePlausible chemistry, likeable data, and studies too small to bet a season on.Underpoweredsupplement
58
40
30
28
36
TE-09Red-light therapyThe mechanism is real, the dosimetry literature is a mess, and your panel is probably not delivering a studied dose.Underpowereddevice
70
35
30
22
34
TE-03Branched-chain amino acidsYou cannot build a twenty-amino-acid protein out of three amino acids, and the trials say precisely that.Unsupportedsupplement
25
70
8
5
31
TE-06Whole-body cryotherapyThree minutes in cold air does not cool your muscle. It numbs you, and it charges forty dollars for the privilege.Overstatedmodality
30
35
22
8
24
TE-07Cupping therapyA controlled bruise with an excellent story. The perceptual effects are real; the stated mechanism is not.Unsupportedmanual therapy
20
25
18
25
22
MECH
Mechanistic plausibility. Does the stated mechanism survive physics, chemistry, and dose? A story about fascia has to beat fascia's tensile modulus.
EVID
Evidence quality. How good the studies are, not how favourable. A well-run trial that finds nothing scores high here: BCAAs are well studied and consistently null.
EFF
Effect size. How large the surviving effect is on outcomes an athlete or a coach could actually measure.
VAL
Cost per unit benefit. Benefit per dollar, including what the dollar could have bought instead. Sleep and food set the bar.

A high evidence score is not praise. BCAAs score 70 on evidence quality because they have been studied carefully and repeatedly. The studies are good. The result is nothing.

Low is not the same as negative. Underpowered means the literature cannot answer the question yet, usually because of tiny samples or unreported dosing. That is a different failure from a broken mechanism.

Cost is an axis on purpose. A real two percent effect for forty dollars a session competes against sleep, food, and training, all of which are free and beat everything in this table.

The short version

If you only read two lists

Ranked by the same published composite used in the matrix. No editorial thumb on the scale, and no affiliate links anywhere on this platform.

Worth the money

Buy these first

High evidence quality, an effect you could measure, and a price that does not require a story.

  1. 01Creatine monohydrateThe most tested ergogenic aid in sports nutrition, and it costs about a dime a day.Supported87
  2. 02Beta-alanineSupported, and narrow: if your effort does not last between one and four minutes, this is not your supplement.Supported72
  3. 03Ice baths after liftingExcellent for playing again tomorrow. Measurably bad for getting bigger.Conditional69
  4. 04NSAIDs for training sorenessThey work on the pain. That is the problem: the pain pathway is also the remodelling pathway.Conditional62

net = 0.35 evidence + 0.3 effect + 0.25 value + 0.1 mechanism

Not worth the money

Skip these, or buy them honestly

Either the mechanism does not hold, or the surviving effect is far smaller than the price implies.

  1. 01Cupping therapyA controlled bruise with an excellent story. The perceptual effects are real; the stated mechanism is not.Unsupported22
  2. 02Whole-body cryotherapyThree minutes in cold air does not cool your muscle. It numbs you, and it charges forty dollars for the privilege.Overstated24
  3. 03Branched-chain amino acidsYou cannot build a twenty-amino-acid protein out of three amino acids, and the trials say precisely that.Unsupported31
  4. 04Red-light therapyThe mechanism is real, the dosimetry literature is a mess, and your panel is probably not delivering a studied dose.Underpowered34

net = 0.35 evidence + 0.3 effect + 0.25 value + 0.1 mechanism

Where to go next

A verdict is the beginning of the explanation

Science LibraryThe mechanisms these verdicts keep referring back toOpenProtocol LabMove the variables and watch the physiology answerOpenRecovery AtlasTraditions on their own terms, then on the evidenceOpen

Education, not diagnosis. This is a student-authored science platform. Nothing here replaces a physician, a physical therapist, or an athletic trainer. Sudden severe pain, numbness, an inability to bear weight, or visible deformity means stop reading and get seen.