THE OCCABUZZ STANDARD
A rating is a claim about evidence, not enthusiasm.
Most of this field sells narrative — "private", "elite", "vetted" — and asks you to trust the tone. We publish the opposite: the exact rule by which every grade is assigned, the formula that resolves it, a worked example you can check, and the one policy that keeps a rating honest. The confidence of our language is bound to the grade — never louder than the proof allows.
Every asset is scored on three independent layers, each 0–100. A thing can have overwhelming evidence and still be useless to you if it does not apply to a healthy operator, or if no one can sustain it. So we grade all three — and refuse to let a strong layer hide a weak one.
Peer-reviewed human data — sample size, effect size, and trial stage. A mechanism that is established, not merely inferred. This is the ceiling: nothing scores above what the trials actually show.
Does it apply to a healthy apex operator, or only to clinical deficit? A drug proven to fix a disease is not proof for the well. The question is always: proven for whom?
Does it survive a demanding calendar? A benefit that requires a protocol no one sustains is a benefit that does not happen. Zero-friction beats theoretically-optimal.
The score is the geometric mean of the three layers, not the average. This is deliberate. An average lets a brilliant marketing story (high Implementation) paper over absent human proof (low Evidence). The geometric mean does not: if any single layer is near zero, the whole score collapses toward zero. A protocol cannot buy back its missing evidence with convenience. A single failed layer collapses the score — by design.
This is not a diagram of the method — it is the method. The three numbers below are the published audit for this dossier; the result is computed by the exact function the whole compendium uses. Nothing is rounded by hand.
apoB causal & actionable — multivariable Mendelian randomization; LDL-C effect reverses to null once apoB is included.
Applies to the healthy operator: this is prevention, measured before disease — not a rescue in the already-sick.
One cheap blood test; Lp(a) measured once for life. The friction is low, but it needs a physician to act on the number.
Demonstrated in humans. Randomized controlled trials or meta-analyses with meaningful sample and effect sizes, and an established mechanism. The strongest grade we publish — and still not a promise.
The instrument is proven, the promise is not. A device or sensor validated against a clinical reference (polysomnography, ECG) — but the health outcome the marketing staples to it is a separate, unproven claim. We grade the hardware, not the hope.
Promising but unfinished. Mechanism is plausible and supported by early human or strong preclinical data, but the proof is still narrow — small trials, unmapped safety windows, effects not yet replicated at scale. Published with its limits stated, never hidden.
Mechanism only. Robust in cells or animals, but human outcome evidence is absent or too thin to act on. Published as a frontier signal for informed operators — explicitly not a protocol.
The invisible grade. Anything that fails a single audit layer never reaches the grid — however loud the marketing behind it. What we omit is as much a part of the method as what we publish.
Every curator will tell you what they look for. Almost none will tell you what they throw away — and the discard rules are where the judgment actually lives. A method that publishes only its inclusions is not a method; it is a taste. So here is the other half, in the order these most often fire.
The trial moved a surrogate — a score, a marker, a self-report — and the marketing swapped it for the outcome you actually want. A device that raises a sleep score has not been shown to make you sleep better; it has been shown to raise its own score. When the endpoint studied is not the endpoint sold, the claim is cut.
The strongest single filter we apply. An intervention that corrects a deficit in a patient is not evidence for an already-optimised, healthy operator. Repletion is not enhancement. If the human data exists only in clinical deficit and is sold to the well, it does not clear Context — however good the trial was.
Funded, designed, analysed and written by the party that profits, with no independent replication. We do not reject industry funding by itself — most of the field would vanish. We reject the closed loop: when every study in a literature traces back to one commercial interest and no one outside it has reproduced the finding.
Statistical significance is a claim about noise, not about your life. A change that clears p<0.05 and moves nothing you would ever notice is a finding, not a protocol. Where an effect is real but trivial, we say so and grade the honest size — we do not round a marginal result up into a directive.
The benefit may be plausible and the mechanism sound, but the dose, the duration, the interactions, or the exit are unknown. An intervention with an unbounded downside does not become publishable because its upside is interesting. It is either graded with the gap stated in the open, or it is held.
The study used a dose, form, route, or purity the product does not deliver. This is the most common failure in the supplement market: real research on a molecule, attached to a package that contains a fraction of it, a different salt, or a route that never reaches the tissue. The evidence belongs to the compound that was tested, not to the label that cites it.
A protocol that demands conditions an operating adult cannot hold is a benefit that does not happen. Implementation is scored as an independent layer for this reason, and because the score is a geometric mean, an intervention that is unsustainable collapses regardless of how strong its evidence is.
- ›Preclinical evidence is not a rejection — it is a grade. Cell and animal work is published as PRECLINICAL, labelled as mechanism-only and explicitly not a protocol. We do not discard the frontier; we mark it, so an informed operator can see the edge of the map instead of being told it is solid ground.
- ›Absence of regulatory approval is not a rejection. Approval answers a different question than evidence does. We grade what the human data shows and state the regulatory status separately, because conflating the two is how both over-claiming and under-reading happen.
- ›Being cheap, unpatentable, or commercially uninteresting is not a rejection — and it is where some of the strongest evidence in the compendium lives. Nothing is graded down for having no margin in it, and nothing is graded up for having plenty.
- ›Disagreeing with consensus is not a rejection, and neither is agreeing with it. The grade follows the evidence and the audit layers. Contrarianism is not a method; it is the same bias running in reverse.
This is the fair criticism of any platform that rates what others sell, so we answer it directly and in public. OCCABUZZ is funded by the operators it serves and by no one else — it sells no products and earns no affiliate commission on anything it grades. Here is exactly what that independence does and does not mean:
- 01A grade is assigned on the evidence alone — before any commercial relationship exists, and it does not move because one begins.
- 02OCCABUZZ sells nothing and earns no affiliate commission on any product, molecule, or device it names — our revenue comes only from the operators we serve, never from what we grade. No purchase or relationship can buy a higher grade, a softer caveat, or a place in the grid.
- 03We are not paid to rate favorably, and a brand cannot purchase its rating. If a graded item later fails, the grade falls — in public, on the record.
Every dossier states, explicitly, where its certainty ends — the exact point past which the evidence stops and speculation would begin. We publish the ceiling, not just the promise. And because the evidence moves, the grades move: each asset carries a revision history and an audit date, and when new data lands, the score is re-cut in public, on the record. A grade that never changes is a grade that stopped reading.
CALIBRATION OVER CERTAINTY.
A standard with no one behind it is a slogan. This one is engineered and audited by a named, accountable operator — Quene Pereira da Silva, Chief Intelligence Officer. The reasoning is on the page, the sources are linked, and the errors, when they come, are corrected in public.
Each asset is scored on three independent 0–100 layers — Evidence (peer-reviewed human data), Context (does it apply to a healthy operator, not only clinical deficit), and Implementation (can it be sustained with zero friction) — then resolved to a calibrated certainty by geometric mean, the cube root of E times C times I. Because it is a geometric mean, a single weak layer collapses the whole score.
No. OCCABUZZ sells nothing and earns no affiliate commission on any product, molecule, or device it names — its revenue comes only from the operators it serves, never from what it grades. A grade is assigned on the evidence alone and does not move for any commercial reason. If a graded item later fails, the grade falls — in public, on the record.
CLINICAL is the strongest grade OCCABUZZ publishes: demonstrated in humans through randomized controlled trials or meta-analyses with meaningful sample and effect sizes, and an established mechanism rather than an inferred one. It is still not a promise — every dossier states exactly where its certainty ends.
It is not published. Anything that fails a single audit layer never reaches the grid, however loud the marketing behind it. The omission is part of the method: what OCCABUZZ leaves out is as deliberate as what it grades.
Seven discard rules, published in full on this page. A claim is cut when the outcome sold was never the outcome measured; when the human evidence exists only in clinical deficit and is marketed to the healthy; when the entire literature is owned end to end by the seller with no independent replication; when the effect is statistically real but too small to matter in a life; when the safety window — dose, duration, interactions, exit — is unmapped; when the product does not deliver the dose, form, or route that was actually studied; or when the protocol cannot survive a working calendar. Because the score is a geometric mean of Evidence, Context and Implementation, failing one layer collapses the whole result.
No. Preclinical evidence is not a rejection at OCCABUZZ — it is a published grade. Work robust in cells or animals but without human outcome data is graded PRECLINICAL, labelled as mechanism-only, and stated explicitly not to be a protocol. Discarding the frontier outright would be simpler and less honest: it would hide the edge of the map rather than mark it. What is rejected is not the preclinical stage, but a preclinical finding sold as though it were a human result.
Yes. Each asset carries a revision history and an audit date. When new evidence lands, the grade is re-cut in public, on the record. A grade that never changes is a grade that stopped reading.
The methodology is engineered and audited by a named, accountable operator — Quene Pereira da Silva, Chief Intelligence Officer of OCCABUZZ. The reasoning is shown on each page, the sources are linked, and corrections are made in public.
The public compendium applies this standard to the field. The Operator Dossier applies it to you.