Every fragrance page shows two numbers dressed up as one clean verdict: a risk band (“Safe bet,” “Some risk,” “Risky”) and a confidence percentage. Here's what's actually behind both, in plain language.
The risk band measures agreement, not quality
The underlying number — polarization score — has nothing to do with whether a fragrance is good. It measures how much reactions to it tend to split. A fragrance almost everyone reacts to similarly, whether they love it or find it merely fine, scores low. A fragrance that reliably produces both passionate fans and firm dislikes scores high. “Risky” is a statement about consensus, not a warning about quality.
Why the cutoffs are set where they are
The line between “Safe bet” and “Some risk,” and between “Some risk” and “Risky,” isn't arbitrary. Both are set from the actual distribution of scores across the full catalog — roughly the 60th and 85th percentile of real polarization scores, not round numbers picked because they looked reasonable. That matters more than it sounds: an arbitrary fixed cutoff can end up never firing for any real fragrance if it doesn't match how scores actually cluster in practice. Grounding the bands in the real distribution means each label reflects where a fragrance actually falls relative to everything else in the catalog, not an abstract scale.
What the confidence percentage actually means
The confidence percentage answers a different question entirely: how much of a fragrance's notes and accords are backed by real, matched data versus a category-level estimate. A high-confidence score means most of what drives the verdict is measured, not guessed. A lower one means more of the score leans on estimating how an unresolved note probably behaves, based on the category it falls into. It's not a claim that every individual ingredient has been lab-verified — it's a signal for how much to trust the number as precise versus directionally useful.
Why some fragrances show “Unknown”
A fragrance with no resolved notes or accords yet would otherwise land at exactly the same raw score as a genuinely low-risk, well-loved fragrance — zero. That's a meaningfully different situation (no data, versus real agreement), so it's treated separately and labeled “Unknown” rather than shown as a false “Safe bet.” If you land on a fragrance page and see “Unknown,” that's exactly what it means — not enough resolved data yet, not a hidden risk.
Why this beats a note list
A note list tells you what went into a fragrance. It doesn't tell you how people who've actually worn it responded, or how much you should trust that response as representative. That's the entire gap this score is built to close — and why the honest version of the answer includes how confident the number itself is, not just what the number says.