How to read the evidence
Every claim here carries two independent marks. One says how well supported it is. The other says what kind of thing it is. They are separate on purpose: a well-supported guess is still a guess, and nothing on this platform can borrow the appearance of a finding by scoring well.
Finding or hypothesis
This is the distinction that matters most, so it is the one carried by shape — the thing you can see without reading.
- FindingDocumented
- Reported directly by a cited source. The grade describes how strong that evidence is.
- HypothesisInferred
- Assembled from separate findings connected through shared biology. No source reports this claim directly.
A hypothesis is never a safety clearance, and neither is silence. If this platform shows no connection between two compounds, that means nothing has been curated — not that nothing exists.
How strong the evidence is
Read the shape first — continuous or broken — then the weight. The grade is computed from recorded facts about the studies behind a claim, and nobody types it in. What decides a grade.
| Grade | Means | Finding | Hypothesis |
|---|---|---|---|
| Well establishedGrade A, Well established. Replicated human trial evidence, consistent direction. | Replicated human trial evidence, consistent direction | ||
| LikelyGrade B, Likely. Human evidence, limited replication or design limits. | Human evidence, limited replication or design limits | ||
| SuggestiveGrade C, Suggestive. Weak human evidence, or strong evidence in a mismatched population. | Weak human evidence, or strong evidence in a mismatched population | ||
| PreliminaryGrade D, Preliminary. Non-human or in vitro only. | Non-human or in vitro only | ||
| SpeculativeGrade F, Speculative. Mechanistic inference, or no evidence found. | Mechanistic inference, or no evidence found |
Chains get weaker
When a connection runs through several steps, its grade is the weakest step, dropped one level for every step past the second. Three well-established links do not make a well-established conclusion.
This is deliberate, and it is the defence against the failure that defines this whole category of content: X activates Y, Y is involved in Z, therefore X causes Z. Three real findings chained into a claim nobody made. Under this rule that chain renders as weak, visibly, without anyone having to catch it.
- Well establishedGrade A, Well established. Replicated human trial evidence, consistent direction.one step, all well established
- Well establishedGrade A, Well established. Replicated human trial evidence, consistent direction.two steps, all well established
- LikelyGrade B, Likely. Human evidence, limited replication or design limits.three steps, all well established
- PreliminaryGrade D, Preliminary. Non-human or in vitro only.five steps, all well established