Methodology & Data Sources
How every number on this site is sourced, what confidence it carries, who checks it, and what changed — the public counterpart to the per-chart attribution.
Last methodology update September 8, 2026
Dataset rules
The six rules no number on this site is allowed to break.
- Two independent sources before a system ships. A sizing system enters the verified dataset only when at least two independent sources agree on its defining relationship — for example, Wikipedia plus Larson Jewelers for the US and UK scales, or two Turkish jewelers for the Turkey formula.
- Definitional relationships beat copied tables. Where a standard defines the size directly (EU size equals inner circumference per ISO 8653:2016; German size equals inner diameter), values are computed from the dataset’s own millimeter measurements rather than transcribed from a third-party table.
- Sourced irregularities are preserved, not smoothed. When a corroborated source contains a historical quirk — such as the UK "Q¼" at US 8.5 instead of a clean half-letter — it is carried through faithfully. Altering it would mean inventing a value no source contains.
- One disclosed interpolation, nothing hidden. A single cell (Japan at US 3.5) has no direct source row and was linearly interpolated between sourced neighbors. It is disclosed in the research log; every other cell is directly sourced.
- Exclude rather than guess. Systems that cannot clear the verification bar stay out: the generic India column remains null (guarded by a validator that fails the build if it is ever populated), and researched-but-unverifiable country systems are documented as excluded with reasons.
- Mirrors are labeled as mirrors. Brand and country tables that restate one external guide — manufacturer, retailer, or third-party — say so on the page with the source name, link, review date, and origin, instead of borrowing the verified dataset’s authority.
Core dataset scope (v1): US/Canada sizes 3.0–13.0 in half-size steps, with UK/Australia, EU/ISO, Germany/Netherlands, Turkey, Japan (moderate confidence), and four named India retailer conventions. Extending the range is a scope decision for a later phase, not a data-accuracy one.
Primary sources
Standards, corroborating retailers, and every mirrored table source — with honest access notes where a standard could not be fetched directly.
| Source | Used for | Access |
|---|---|---|
| Wikipedia, "Ring size" article | System descriptions, standards citations, and the raw equivalency table behind US, UK, and Japan values. | Fetched directly. Carries its own original-research banner, which is why single-source values from it stay moderate-confidence or excluded. |
| US Bureau of Standards, Circular No. 43 (1921) | Definitional relationship behind the US/Canada scale: whole-size step of 0.032 in (0.8128 mm) inner diameter. | Cited via Wikipedia; not freely available in independently re-verifiable form. |
| BS 6820:1987 (British Standard) | Definitional relationship behind the UK scale: one letter step equals 1.25 mm of inner circumference. | Cited via Wikipedia; paywalled standard, not fetched directly. |
| ISO 8653:2016 (Jewellery — Ring-sizes) | Definitional relationship behind the EU scale: size number equals inner circumference in millimeters. | Cited via Wikipedia; paywalled standard, not fetched directly. |
| Larson Jewelers, International Ring Size Conversion | Independent corroboration for US, UK, EU, and Japan descriptions, plus the Japan-system formula used as a cross-check. | Retailer site, consulted directly during research passes. |
| Taj Amsterdam, ring size assistance | Third independent worked example confirming the German/Dutch diameter-based system. | Real Dutch jewelry retailer, fetched directly. |
| Bluestone and Jewelove published size guides | Two retailers independently publishing the identical circumference-minus-40 formula — the strongest India corroboration found. | Each guide read directly against its own published table. |
| Samsung Electronics — Galaxy Ring Sizing & Fit Guide | Mirrored table restated verbatim on its chart page with per-page attribution. | Last reviewed 2026-08-25; re-check the live guide before ordering. |
| RingConn Official — Gen 2 Sizing & Fit Guide | Mirrored table restated verbatim on its chart page with per-page attribution. | Last reviewed 2026-09-08; re-check the live guide before ordering. |
| Ultrahuman Official Sizing Specifications | Mirrored table restated verbatim on its chart page with per-page attribution. | Last reviewed 2026-09-08; re-check the live guide before ordering. |
| Movano Health — Evie Ring Sizing & Fit Guide | Mirrored table restated verbatim on its chart page with per-page attribution. | Last reviewed 2026-09-08; re-check the live guide before ordering. |
| Pandora official ring size guide | Mirrored table restated verbatim on its chart page with per-page attribution. | Last reviewed 2026-08-25; re-check the live guide before ordering. |
| Frank & co. | Mirrored table restated verbatim on its chart page with per-page attribution. | Last reviewed 2026-08-25; re-check the live guide before ordering. |
| ringssizechart.com | Mirrored table restated verbatim on its chart page with per-page attribution. | Last reviewed 2026-08-22; re-check the live guide before ordering. |
| ringssizechart.com | Mirrored table restated verbatim on its chart page with per-page attribution. | Last reviewed 2026-08-22; re-check the live guide before ordering. |
| ringssizechart.com | Mirrored table restated verbatim on its chart page with per-page attribution. | Last reviewed 2026-08-24; re-check the live guide before ordering. |
| ringssizechart.com | Mirrored table restated verbatim on its chart page with per-page attribution. | Last reviewed 2026-08-24; re-check the live guide before ordering. |
Smart-ring model and version history
One row per tracked model and generation, with the verification date behind each.
| Model | Versions | Sizes | Table origin | Last verified | Model guide |
|---|---|---|---|---|---|
| Oura Ring Gen3 | Gen3 | 6–13 | Third party | 2026-08-22 | Open guide → |
| Oura Ring 4 | Ring 4 | 6–13 | Third party | 2026-08-22 | Open guide → |
| Samsung Galaxy Ring | Galaxy Ring | 5–13 | Manufacturer | 2026-08-25 | Open guide → |
| RingConn Gen 2 | Gen 2 | 6–14 | Manufacturer | 2026-09-08 | Open guide → |
| Ultrahuman Ring AIR | Ring AIR | 5–14 | Manufacturer | 2026-09-08 | Open guide → |
| Evie Ring | Evie Ring | 5–12 | Manufacturer | 2026-09-08 | Open guide → |
Compare models side by side on the Smart Ring Sizing by Model hub.
Reviewers and review process
No invented expert panel — the process, stated plainly.
- Source-based verification. Every value traces to a documented source in the research log, with conflicts and their resolutions written down — including cases where the losing finding (such as a conflicting retailer chart) is what set a confidence level.
- Automated dataset validators. Unit-test suites re-check dataset sort order, circular-measurement consistency, source-attribution shape, and quality gates on every test run, and the build refuses to ship if the generic India field is ever populated.
- Build-time site validations. Every build validates unique titles and descriptions, self-referencing canonicals, structured-data presence, internal-link integrity with no orphaned indexable pages, and reciprocal hreflang annotations.
- Public correction channel. Readers who spot a stale figure can report it via the contact page; confirmed corrections update the affected row, move its review date, and append to the log above.
Correction log
Dated data changes, newest first. Fixes stay visible instead of silent.
· Smart-ring tables
Added manufacturer-specification tables for RingConn Gen 2, Ultrahuman Ring AIR, and Evie Ring, each attributed with its official sizing guide and review date.
· Turkey system
Added the Turkey column after two Turkish jewelers independently published the same circumference-minus-40 relationship; values derive directly from that formula.
· Japan confidence
Rechecked Japan values against a current retailer chart, found differences at several rows, and kept the moderate-confidence classification instead of claiming either mapping as universal.
· Russia research
Folded Russia into the Germany field’s documentation after finding a matching worked example, instead of adding a redundant column with identical numbers.
· Brazil / Swiss pages
Published Brazil and Swiss references explicitly as labeled third-party material with a documented cross-check pass, rather than as verified national systems.
Found something stale? Report it via the contact page with the page URL and what the official guide currently says.
Frequently asked questions
Where do the ring size numbers actually come from?
The core table combines definitional standards (ISO 8653 for EU sizes, the US Bureau of Standards diameter relationship, the British Standards letter step) with corroborating retailer tables such as Larson Jewelers and Taj Amsterdam. Every system needs at least two independent sources before it ships, and the full paper trail lives in the research log linked from the sources section.
What does "moderate confidence" mean for Japan?
Two sources corroborate the Japan scale but drift by a size or two at the top end, and one cell had to be interpolated between sourced neighbors. That is disclosed rather than smoothed over: Japan works for general conversion guidance, but confirm with the retailer for high-value pieces instead of treating it as exact.
Why is there no single India size on this site?
Because none exists. Major Indian retailers publish at least four mutually incompatible numbering conventions under the name "India ring size," so a single India column would mislead more than it helps. The site exposes four named retailer conventions instead, and a build-time validator fails the build if anyone ever populates a generic India field.
How are mirrored brand tables different from the core dataset?
Mirrored tables restate one external guide verbatim — a manufacturer spec, one retailer chart, or a labeled third-party reference — and say so on the page with source, link, and review date. They deliberately do not borrow the core dataset’s verified status, which is why the Oura rows, for example, are flagged third-party while Samsung rows cite Samsung directly.
How often is sizing data re-verified?
Every mirrored table carries its own last-reviewed date, shown on its chart page and in the model history below. Manufacturer revisions, new sizes, or new generations move the affected row’s date; the correction log records what changed and when.
I spotted a stale or wrong figure — what happens if I report it?
Send it via the contact page with the page URL and what the official guide currently says. Confirmed corrections update the affected row, move its review date, and append to the public correction log below — the log exists so fixes stay visible instead of silent.