Research · White paper
MPAE — the archetype model behind SailrScience.
A measured, parameterized decomposition of how offshore boats actually vary — and the structural reason FleetEdge™ surfaces 11 archetypes rather than a marketing taxonomy.
Photography: FleetEdge™ / SailrScience archive.
What MPAE is
MPAE is a decomposition. Each ORC certificate carries hundreds of measured parameters — geometry, sail plan, stability profile, predicted performance. Most of those parameters are correlated; many are derivatives or ratios of a smaller number of independent measurements. The model identifies the small number of independent axes along which boats actually differ, projects every measured boat onto those axes, and observes that the projected boats cluster.
The clusters are the archetypes. Eleven of them, in the current model. They were not designed top-down; they were discovered.
The archetype boundaries are not opinions. They are properties of the measured data, and they have held up under cross-validation, across regions, across class systems, and across more than four years of new certificates issued since the model was first trained.
What MPAE does
Two things, primarily.
It places every boat. Given a certificate, MPAE returns the archetype the boat sits in and the distance from the cluster center. The placement is decision-grade: it predicts which other boats are the right comparison cohort for analytics, for racing, for handicap fairness studies.
It conditions every analytic. A fleet-wide read of crew performance is meaningless without the archetype axis — you cannot compare a heavy-displacement sleigh ride to a planing form to a stripped-out flyer and learn anything about the crews. MPAE supplies the structure that makes the comparison possible.
Everything FleetEdge surfaces at the fleet scale — dimension reads, race-cohort analysis, weather-conditioned reads — is archetype-conditioned underneath. The model isn’t visible on the surface. It’s what makes the surface mean anything.
White paper
The full white paper formalizing MPAE is authored in the research arm. It documents the parameter set, the feature engineering, the decomposition method, the cross-validation results, and the cluster definitions.
The white paper is available on request to institutional contacts, naval architecture firms, partner sail lofts, and serious racing programs.
PDF download will be wired here when the next revision is publicly released.
Where MPAE shows up in product
The 11 archetypes that MPAE produces are the structural backbone of the FleetEdge analytics surface. The archetype map, the dimension decomposition, and the cohort-conditioned race analysis are all surfaces over the same model.
Discuss the model
For researchers, naval architects, journalists, and institutional contacts: the research arm responds to questions about the model and the methodology.