Studies & Research

The pages that describe the scoring engine, the evidence behind it, and the stand-alone research notes.

Methodology

How We Score Safety — The scoring pipeline, factor by factor, with the figures of the latest weekly run
Scoring White Paper — Complete mathematical specification of the engine

Empirical Validation Study — How the engine was built from FMCSA data and how it performs out of sample

Stand-alone studies on public FMCSA data. They inform the engine where the page says so; none of them is the model behind the grade.

Insurance Market Census — Who insures motor carriers today, by line of business, fleet size and state, from the insurance filed with FMCSA (updated weekly)
The Inspection Paradox — What an inspection-based score measures, tested at the border
Violation Type Study — Which violation categories go with crash risk (February 2026)
Crash Model — Machine Learning — A truck-level machine-learning experiment, technical write-up (February 2026)
ML Crash Model — The same experiment in plain language (February 2026)

This page is itself an orphan — not linked from navigation.