Methodology and software limits

We use public FMCSA Safety Measurement System input files, including inspections, violations, the carrier census and crashes. The monthly snapshots are the source for the calculation; uploaded documents are handled separately.

We record the checksum, row count and schema of each downloaded file and retain the raw file unchanged. Corrections are applied to derived data, so a calculation can be traced back to its source.

We use FMCSA’s SMS Methodology, version 3.21. BASIC stands for Behavior Analysis and Safety Improvement Category.

Within each inspection and each BASIC, applicable violation severity weights are added together, including out-of-service additions where required. The sum is capped at 30 before the inspection’s time weight is applied. Violations in other BASICs do not count toward that cap.

For inspection-normalized BASICs, weighted violation points are divided by the total time weight of relevant inspections, including those without a violation in that BASIC. Other BASICs use different exposure denominators. A violation’s contribution is the difference between the measure with and without that violation, accounting for the cap.

A calculated value and its displayed precision are different. Rounding and truncation can produce different displayed values, so comparisons must account for the precision used in the published figure.

A displayed match alone does not establish a system’s internal arithmetic. A reproducible comparison needs the source snapshot, calculation rules, display precision and carrier population.

A severity weight is an input to the calculation. Weighted numerator contribution accounts for time weighting and the inspection cap. The marginal change in a named BASIC measure also depends on its denominator. A percentile compares a BASIC measure with those of carriers in a similar safety event group. Do not add points across BASICs. These SMS calculations are separate from an official safety rating.

For property carriers, relative percentiles and alerts are withheld from the public SMS display. Public measures are available for Unsafe Driving, HOS Compliance, Driver Fitness, Controlled Substances/Alcohol and Vehicle Maintenance; Crash Indicator and HM Compliance results are not public. Carriers can sign in through their own FMCSA Portal account to view complete SMS results (SMS Methodology p. 1-2, note 6; FMCSA FAST Act guidance).

This is our estimate from public FMCSA data, not an official FMCSA percentile.

Published passenger-carrier measures provide one comparison set for reconstructed measures. Matching a measure does not establish the accuracy of an estimated percentile, and results for one carrier population do not establish accuracy for another.

An accuracy comparison needs a named snapshot, the BASICs and carriers included, and the errors observed. The largest error in that sample does not bound errors for every carrier or future snapshot.

This is our estimate from public FMCSA data, not an official FMCSA percentile. Read the estimated position together with its range and the snapshot date shown in the report. If the interval crosses the intervention threshold, the estimate cannot put the carrier firmly on one side of the line.

Every report names the snapshot it was computed from. If new files have not arrived or fail the import checks, we continue to show the previous snapshot with its date.

Import checks cover file structure, completeness and integrity. A failed import does not become a new report snapshot.

If there are no relevant inspections in the calculation window, we state that the data is insufficient. An absence of inspection data is different from an inspected record with no applicable violations.

If the measure can be calculated but an estimated position does not pass its accuracy check, we show the measure and violation contribution and explain why the position is unavailable.

Unsafe Driving is excluded from the free lookup: no records, full measure or estimated position are offered. The engine’s internal numerator points are not a full measure. Its SMS denominator is average power units × utilization factor, with VMT and carrier-segment inputs (Equation 3–1). Our product lacks the historical power-unit inputs needed to reconstruct and validate it. The coverage details below list all seven BASICs and explain availability.

Uploaded documents are held separately from the public data used to calculate measures, in encrypted storage. Only values you confirm can enter the draft.

Files are deleted 14 days after upload; uploading the same file again does not extend that period. The case page and confirmations expire no later than 14 days after your first upload.

See the Privacy policy for processing providers, backups and deletion requests.