How the map was built

Every rated link traces to at least one published source. Nothing is inferred and nothing is invented.

WHAT CHANGED IN THIS VERSION

The previous map allowed each rated link exactly ONE source. That was a limit of the spreadsheet, not of the research — and it was hiding things. Many links rested on two or three studies; the extra ones were named in the text but could never be reached.

The map now holds ONE ROW PER SOURCE. 1,133 rated links, 1,254 evidence links drawn from 558 distinct documents, every one clickable. 117 links carry more than one.

WHAT WAS FOUND AND FIXED

Every link in the map was opened and checked. Every claim was tested against the ABSTRACT of the paper it cites — not its title, not its keywords.

Seven links pointed at the WRONG PAPER. A cell on iron and anaemia cited a paper about the unusual causes of anaemia. A cell on smoking and lung cancer cited a paper on air pollution. A cell on waist-to-height and diabetes cited a coffee trial. A cell on body weight and blood pressure cited a paper on slow breathing. All seven are corrected.

28 sources came from places that do not meet the evidence standard — a predatory publisher, a clothing shop, a content farm, an AI search engine, a supplement vendor, blogs, and several dead pages. All were removed and replaced. They are listed in sheet 10, so the record survives even though the links are gone.

HOW STRENGTH IS GRADED

Strength grades the EVIDENCE, not the size of the effect. Where a link rests on several studies, it is graded on the STRONGEST one.

●●● randomised trials, meta-analyses, systematic reviews, clinical guidelines, or an established clinical syndrome

●● cohort or observational studies with some trial support

● early or single-study signals

Six links are graded ●●● as deliberate, documented exceptions to that rule. They are marked as such in the research catalog, so they can be seen and argued with rather than mistaken for rule-following.

LIMITS — read before acting on anything here

1. A link is not a cause. Most ●● and ● rows are associations, and associations mislead.

2. A blank means no established link IN THIS MAP — not proof none exists. Evidence is uneven.

3. Deficiency is far better studied than excess. The 'don't overdo' side is thinner by nature.

4. The map says nothing about YOU. It says what has been found across populations.

5. Fixing a shortfall helps only if you HAVE the shortfall. More of a nutrient you already have

enough of buys nothing — and past the ceiling in sheet 4, starts costing you.

6. An abstract is not a full paper. Claims were checked against abstracts; a figure buried in a results table could not be.

Two nutrients (Linoleic acid, Starch) are tracked but carry no rated link yet.

Alcohol and tobacco appear only as things to cut. They are never levers to add.