Editorial Standards

Last updated: 3 October 2026

StrengthInsight is a research publication, not a laboratory. This page explains what that means in practice: how we research a guide, how we evaluate software and open-source projects, what we are willing to claim, and what we deliberately refuse to claim.

Why we publish this

Fitness wearable content is unusually polluted. A large share of articles comparing trackers are written by people who own none of them, quote no sources, and describe testing they never did. Readers cannot tell the difference from the outside, so we think the honest approach is to state our method up front rather than ask you to infer it.

We are a small independent site with no institutional affiliation and no commercial relationship with any manufacturer we cover. We tell you that plainly, and you should weigh it when you read us.

How we research

Every guide on this site is compiled from sources we can point you to. In practice that means, in order of preference:

Vendor documentation and support articles. What a company says its own metric means is the primary source for that metric. We treat vendor claims as claims, not findings.
Published research. Where we cite a study or a stated methodology, we name it and link it. We do not cite a finding we cannot locate.
Independent testing and teardowns published by other outlets, cited by name.
Arithmetic. A lot of what we explain — estimated one-rep max, plateau thresholds, volume maths — is derived from first principles rather than observed. We show the working.

Where a guide relies on a source, it says so in the text and lists it under Sources at the foot of the page.

How we evaluate software, apps and open-source projects

A growing part of this site covers software rather than hardware: dashboards, API clients, data bridges and open-source projects. That genre carries a specific risk, because a list of tools is easy to produce and easy to get wrong. So we publish the rubric.

What we read

Everything we assert about a tool comes from material the tool's own authors published:

Official vendor documentation and API references — for what an API returns, what a subscription unlocks, and what a platform has retired.
The project's own README and documentation — for install steps, supported data sources, and stated limitations.
The licence file — because "free" and "you may use this commercially" are different things. PolyForm Noncommercial, AGPL and MIT have materially different consequences for a reader, and we name which one applies.
Commit and release history — to judge whether a project is maintained or abandoned. Several projects we cover are archived, and we say so.
Release notes and changelogs — because platform behaviour moves faster than documentation. Several times the documentation was still describing a limitation that had already shipped.

What we have not done

We have not installed or run the software we write about. The comparisons in our tools and dashboards coverage are desk research: documentation, licences and history, read carefully and cited. We did not stand these projects up, connect a wearable, or verify a sync end to end. Where a guide describes how something behaves, that description is derived from its documentation, not from our own observation.

We publish that limitation rather than hiding it, for two reasons. The first is honesty: you are entitled to know how a claim was produced before you decide what weight to give it. The second is practical. A guide written from documentation can tell you what a project claims to do, what it costs, and what it will break on. It cannot tell you whether the sync is reliable in practice, and we would rather be explicit about that boundary than imply we know.

How we handle a project's own claims

A README is a marketing document. When a project says its recovery score is "physiologically accurate", that is a claim by the project's authors, and we report it as their claim. Where a score is computed from a method we can locate in published literature, we name the method and say the implementation is an approximation of that method rather than a reproduction of any proprietary model. Where the formula is simply not published anywhere — which is the normal case for commercial wearable scores — we say that the formula is unavailable and that any reimplementation is a different instrument, not a copy.

What this means for corrections

Desk research has a predictable failure mode: a project ships a breaking change, or a licence changes, and our page becomes quietly wrong. We accept that, and we handle it the same way as any other error — corrected on the page, in public, dated. If you are reading a tools guide and the project has moved on since, treat the specifics as a dated snapshot and check the project's own documentation for current behaviour.

What we do not claim

We do not wear devices for a fixed test period and report the results as our own. If a guide describes how a tracker behaves, it is describing documented behaviour, published testing, or our own reasoning — and it says which.

We specifically avoid three things that are common in this category:

"We tested it for 90 days." Unless it happened and we can show the method, we do not say it.
Manufactured precision. We will not present an illustrative number as though it were a measured result. If we walk through a worked example to show how a calculation works, we label it as an example.
Clinical framing. Nothing here is medical advice. Wearable recovery and readiness scores are not clinical measurements, and we will not present them as diagnostics.

We also avoid unnamed hedge language. If the honest answer is that a device is well regarded without a strong methodology behind the claim, we say that the claim is weak rather than repeating it.

Corrections and thresholds we invented

Some of what we explain is our own construction — a plateau definition, a sample-size rule, a chart convention. That is fine, but it has to be marked. When we use a threshold we chose, we say so and give the reasoning, so you can disagree with the threshold rather than with an unnamed authority.

For example, where we discuss spotting a strength plateau we use a ±3% band across consecutive sessions. That band is our convention for filtering noise, not a published standard, and we label it as such.

Corrections

If we get something wrong we want to fix it and say we did. Corrections are made on the page, not quietly. Material corrections are noted at the top of the article with the date. If a correction changes a conclusion, we say so explicitly rather than editing the text as though it was never wrong.

Write to support@strengthinsight.app with the URL and what you believe is incorrect. If we're right and you're wrong, we'll tell you why — that's a useful outcome too.

Affiliates and how we make money

Some guides contain affiliate links. When they do, the guide says so near the top, before you reach any link, and every affiliate link is marked. You should never have to read a guide to discover it contains commercial links.

Affiliate revenue does not determine what we cover or what we conclude. We do not accept payment for coverage, and we do not review a product we would not write about without an affiliate link. Full detail is on our disclosure page.

Our coverage of software, dashboards and open-source projects carries no affiliate links at all, and the commercial status of each tool is stated in the guide rather than monetised. Where we list a paid service we say plainly that it costs money, what the free alternative is, and what you give up by paying — because a reader choosing between a $5 VPS and a $5-per-month subscription deserves that comparison more than they deserve a referral fee.

What "StrengthInsight" is

The site began as a product — a tool that read workout screenshots and built progression charts. That tool has been retired. What remains is the publication, and it is deliberately narrower: we cover what fitness wearables measure, how their metrics are derived, and how to read your own training data honestly. We would rather be obviously useful about a small number of questions than broadly vague about many.

Affiliations

StrengthInsight is independent. It is not affiliated with, endorsed by, or sponsored by WHOOP, Fitbit, Garmin, Apple, Oura, Hevy, Strava, Amazon, or any other company whose products we cover. Product names and trademarks belong to their respective owners and are used here to identify the products being discussed.