An evidence audit · 26 claims graded
Across 60 meta-analyses of diet and inflammation, 2% reached high certainty. The other 98% is sold at full confidence.
This publication checks the most repeated claims in the category against the trials, grades each one, and shows the source that settles it — including the claims the evidence does not support.
Six short parts · no email needed
Before anything else
What the small squares mean
Every claim on this site carries one. The letter is how strong the evidence is — not how big the effect is, and not how often the claim gets repeated. Darker and heavier means stronger.
- A
Strong evidence
Several independent reviews agree.
- B
Moderate evidence
One good meta-analysis or consistent trials.
- C
Weak evidence
Only observational or lab evidence.
- D
Very weak or untested
Weak, contradicted, or never tested in people.
Tested: no effect
Good trials looked, and the effect was not there.
The free story
Six parts, in the order the argument runs
01
The gap
2% of the evidence is certain. The other 98% is sold as if it were.
Sixty meta-analyses of diet and inflammatory markers were rated for certainty. Two per cent reached high. That one number is the reason this publication exists.
02
The marks
What the small squares mean
Every claim here carries a square with a letter in it. The darker the square, the stronger the evidence. Thirty seconds to learn, and it changes how every health headline reads.
03
The claims
Six popular claims, checked against the trials
Green tea, seed oils, nightshades, leaky gut, the Nordic diet, the seven-day reset. Two were tested properly and did nothing. None of the six holds up as stated.
04
The levers
Diet is the fourth lever, not the first
Ranked by the effect actually measured on inflammatory markers, food comes after fat mass, aerobic exercise and sleep. Nobody selling food lists can say that.
05
The test
Why a single blood test cannot show that anything worked
In a person whose body has not changed at all, repeat hs-CRP results scatter by tens of per cent. The before-and-after test sold across this category measures that scatter.
06
What is left
The short list that survives review
After the claims fail and the levers are ranked, one dietary finding holds at the top grade. It is shorter than any food list for sale, and that is the finding.
Limits
What this does not do
- It makes no claim that any food, pattern or supplement treats, prevents or reduces the risk of disease. That claim is not permitted for food in the EU, and the evidence would not support it here in any case.
- It never states that something “reduces inflammation”. Trials measure named markers, those markers disagree with each other, and the text reports what was measured, in whom, and how certain the synthesis was.
- It promises no result on a timeline. Chapter 4 is the arithmetic for why a personal before-and-after test cannot demonstrate one.
- It contains no elimination phase. No systematic review in this field has ever tested one, and chapter 6 explains what replaces it.
- It is not written for everyone. Chapter 0 opens with ten exclusion criteria, and it is free to read in full.
- It is not medical advice and does not substitute for clinical care.
Contents
Nine chapters. Every one opens in public.
- 00Who this is not forFree in full
- 01How to read thisRead the free opening
- 02The four levers, in orderRead the free opening
- 03What the evidence does not supportRead the free opening
- 04Why a single test result means nothingRead the free opening
- 05What holds upRead the free opening
- 06Addition, not eliminationRead the free opening
- 07ToolsRead the free opening
- 08Bibliography, contradictions and gapsRead the free opening
Access
One payment. The full audit, and every revision.
- All twenty-six claims graded, each with the source that settles it
- The four levers ranked by measured effect, with the arithmetic
- The threshold a test result has to cross before it means anything
- The complete short list of what survives review
- Doses and timeframes the trials used, and five working sheets
$39
USD · one payment · no subscription
- All nine chapters and all 26 graded claims
- Web reader and a downloadable PDF
- Every future correction and revision, included
Tick both boxes above to enable checkout.
Corrections are made in place and logged. Where a grade or a safety statement changes, everyone with access is emailed. There is no subscription, no sponsorship and no affiliate revenue anywhere in the graded text. Pricing is adjusted by region; the price shown is the rate for US, and the amount charged is confirmed at checkout. How the grading works.