Trust, Data & Independence
The commitments behind every FleetEdge figure — sources, boundaries, and the rules that keep the analysis honest.
Every fleet has a fingerprint.
Every figure names its cohort, its denominator, and its date — the vocabulary is public.
Hidden Shape52 dimensions distilled into 11 performance archetypes.
How the corpus clusters before the results arrive.
Read the shape of the fleet before the scoreboard reads it.
SailrScience is an independent analytics company. FleetEdge explains performance outcomes around rating systems — it does not rate boats, issue certificates, alter handicaps, or infer proprietary rule inputs. That sentence governs every ORC-facing and IRC-facing surface on this site, and this page collects the commitments that back it up.
Trust in a fleet-analytics platform rests on four things: knowing where the data comes from, knowing where the analysis stops, knowing that commercial relationships cannot bend the outputs, and being able to check all of it. Each has its own section below, and each links to the canonical page where the full policy lives.
Where the data comes from — and where the analysis stops.
ORC certificate data
Public certificate information — hull, rig, and stability measurements. FleetEdge consumes what ORC publishes; it has no privileged access and no role in rating. The full relationship is documented on ORC & SailrScience.
IRC published results
Finishing positions, elapsed and corrected times from public results — plus published hull dimensions used to match each IRC boat to its ORC design family. FleetEdge never reads IRC rating values, TCCs, or proprietary handicap data. Boundaries: IRC & SailrScience.
Weather & ocean reanalysis
Atmospheric context from the ERA5 reanalysis (~31 km grid) and ocean currents from HYCOM (1/12°). Both pages state what those resolutions can and cannot resolve — see Weather and Ocean Intelligence.
Race results corpus
Published race results from public sources, linked to boats and enriched with conditions. Every count derived from this corpus names its cohort and denominator — the vocabulary is defined on Data Definitions.
Partnerships cannot bend the analysis.
Partner relationships do not influence FleetEdge models, archetype definitions, rankings, or event coverage. Strategic partnerships appear on partner-branded surfaces only — never on country, fleet, event, or archetype pages, and never in any analytical comparison. That separation is permanent, audit-checked, and published with a permalink.
The canonical policy text lives in one place: the Neutrality Policy on the Partners page. If any FleetEdge surface ever appears to contradict it, the policy wins — and we want to hear about it via Contact.
Every page is checked before it can publish.
FleetEdge's public site is built from a single validated data source and passes an automated integrity check on every deployment. The checks are blunt by design: one metric cannot render two different values on a page; every figure must carry its cohort, denominator, and as-of date; an index page and its detail pages must agree; archetype names and shares in narrative text must match the published data; terminology and claims are held to a public-language standard. A page that fails any of these does not publish.
Alongside its scope, every analytical card carries one of five public evidence badges telling you what kind of claim it makes — from live fleet-wide measurement to forward-looking scenario. The closed badge lexicon is documented in How It Works.
The archetype methodology itself is publicly documented by SailrScience Research — overview public, full white paper available on request: the archetype model at SailrScience Research.
The claims themselves are tested in public: see the validation analyses — fairness distribution, crew-effect dispersion, and mapped-IRC prediction accuracy, each with confidence intervals.
Versions, vintages, and what we do not claim.
The taxonomy is versioned. Archetype names and definitions are maintained over time; race and event sections render the view as published nearest the event’s date, while current fleet surfaces render the current taxonomy. The vintage rule — and why an older race section can show a name absent from the current list — is explained on Data Definitions.
Capabilities are labeled. The Roadmap separates what is live today from what is planned; forward-dated capability is never presented as a current measurement.
Projections are projections. A mapped-IRC boat carries a physics profile projected from its ORC design family — a modeled estimate with stated confidence, not a measurement of that hull. Small cohorts fall below the narrative floor and ship structural readouts only, without comparative takeaways. And reanalysis weather describes conditions at model resolution, not at a boat’s masthead.
The relationship pages.
ORC & SailrScience
How FleetEdge works with public ORC data — and the line it never crosses.
IRC & SailrScience
Independence, published-results-only inputs, and the dimension-based bridge.
Partners & Ecosystem
Who we work with, on which surfaces — and the neutrality policy, with permalink.
Data Definitions
The cohort vocabulary behind every count, and the taxonomy vintage rule.
Check us.
The sources are named, the boundaries are published, and the checks run on every deployment.