How we test
Score weights fixed in the template
Every review score on this generator uses the same math. The scale is 0–10. Overall is the weighted sum, rounded to one decimal. A frontmatter file that disagrees with the formula fails the build.
| UV protection | 0.35 |
|---|---|
| Comfort | 0.25 |
| Breathability | 0.25 |
| Packability | 0.15 |
Scores are turned off. No review page is published, and no score number is shown. When a review comes back, a score will be editorial. It is not a textile lab, not a UV-meter session, and not an Amazon rating. Each sub-score reads the features the brand lists for that hat: coverage, comfort of the stated fit and fabric, breathability of mesh or cooling cloth, and how packable the brand says the shape is.
What a score uses
UV protection leans on the coverage the listing names, such as a wide brim, a neck flap, or a stated UPF value, and where that coverage sits. Comfort leans on the size run and the fit notes, including XL, XXL, and adjusters. Breathability leans on cooling fabric, mesh, or a fan the brand describes. Packability leans on whether the listing calls the hat foldable, and on how much brim or flap the shape still has to carry.
The weights are fixed. Overall is 0.35 times UV protection, 0.25 times comfort, 0.25 times breathability, and 0.15 times packability, rounded to one decimal on a 0–10 scale. If the article file disagrees, the build fails. The rating box on a review shows that math.
What we do not record
This site does not publish caliper brim widths, neck-flap lengths, gram weights, UV-meter pairs, soak-and-dry times, or wind-tunnel notes. If a number like that is not in the brand listing, it does not appear here. A later article from the content team can add observations in tested.measurements, but the build does not require a fake lab sheet.
UPF, when a page states one
UPF values are copied from the TOP-EX product page, with the area the brand attaches them to. The Sources list under the article links to that page. The value is not a medical claim and it is not a test this site ran.
Replacing an article
Drop a Markdown file in sites/topex/content/articles/ with the same frontmatter fields the current files use. Product ids have to match sites/topex/site.json. Run npm run build. A missing field, an unknown product, or a review score that breaks the formula stops the build.