CSA Scores and the BASICs

The seven BASICs

FMCSA's Safety Measurement System sorts what happens at roadside into seven Behavior Analysis and Safety Improvement Categories:

Five are public, with a limit many carriers do not know about. Since the FAST Act of 2015, FMCSA does not publicly display percentiles or alerts for property carriers. What anyone — including a broker or shipper — can see for those five categories is your measure and the inspection and violation history behind it. Passenger carriers' information remains fully public, and you see your own percentiles by logging in. That history is still part of your sales pitch whether you like it or not: an out-of-service order is visible even when the percentile is not.

What the number actually means

A BASIC percentile runs from 0 to 100, where 100 is the worst performance and 0 is the best. It is the reverse of a school mark, and people get it backwards constantly.

How a measure is built

Before there is a percentile there is a measure, and the measure is arithmetic you can check against your own record. Under FMCSA's current SMS methodology:

• Each violation has a severity weight from 1 (less severe) to 10 (most severe). In the Hours-of-Service, Vehicle Maintenance, Hazardous Materials and Driver Fitness BASICs, an out-of-service violation adds 2.
• Severity is capped at 30 per inspection per BASIC, so one very bad inspection cannot count without limit.
• Each event has a time weight: 3 if it is under six months old, 2 if six to twelve months, 1 if twelve to twenty-four months. After 24 months it drops out entirely.
• For vehicle-related BASICs, the weighted violations are divided by the time-weighted count of relevant inspections. Clean inspections therefore pull the measure down.

A worked example, with the arithmetic

Illustrative record, and illustrative severity weights chosen for the arithmetic — real weights come from FMCSA's tables. A one-truck carrier has six vehicle inspections in 24 months: two in the last six months, two at six to twelve months, two at twelve to twenty-four months.

1. Inspection time weights: (2 × 3) + (2 × 2) + (2 × 1) = 12.
2. A recent inspection found a defect with severity 6, placed out of service: 6 + 2 = 8, × time weight 3 = 24.
3. An old inspection found a defect with severity 4: 4 × 1 = 4.
4. Measure: (24 + 4) ÷ 12 = 2.33.
5. Add one clean inspection this month: 28 ÷ (12 + 3) = 1.87, a fall of about 20 percent from a single clean result.

And with no new inspections, the recent violation's weight falls from 3 to 2 as it ages past six months, then to 1, then to nothing. That is the arithmetic behind "time repairs a score".

When a percentile appears at all, and when it matters

FMCSA does not calculate a percentile without enough data. It also generally drops carriers whose only violations in a BASIC are more than 12 months old. The minimums and the intervention thresholds:

A carrier above a threshold may be prioritised for intervention: a warning letter, an investigation or closer roadside attention. For a very small carrier, the minimums mean there is often no percentile at all — which is not the same as a good one.

Why small carriers swing so violently

This is the single most useful thing to understand if you run one or two trucks.

Your percentile is built from a small number of inspections. One bad inspection with several violations can move a one-truck carrier from a comfortable position to an alarming one, and a clean year can move it back. A large fleet's score barely notices the same event, because it is diluted across hundreds of inspections.

The practical consequences: do not panic at a single spike, and do not assume a low score means you are being watched more closely than a bigger operator with the same behaviour. The volatility is arithmetic, not judgement.

How the data gets there

SMS takes a monthly snapshot from FMCSA's national database. Roadside inspections, violations and crashes flow in from the states. Changes to your carrier record — a corrected MCS-150, for example — generally show up in the following month's results.

So nothing you do today changes your score today. There is always a lag, in both directions.

What actually moves a score

• Violation severity and recency. Newer events weigh more heavily than old ones.
• Out-of-service violations count for more than paperwork findings.
• Time. Events age out. A clean stretch is the most reliable repair.
• Inspection volume. More clean inspections dilute a bad one — which is why avoiding scales is a worse strategy than passing them.

What to do about a bad BASIC

1. Look at your own record first. Pull your inspections and read the violations rather than the headline number.
2. Challenge genuine errors through DataQs, FMCSA's request system for disputed records. It is worth the effort when a violation is wrongly attributed to you.
3. Fix the cause, not the score. A Vehicle Maintenance problem is a pre-trip inspection problem. See driver vehicle inspection reports and the annual inspection.
4. Expect months, not weeks. Monthly snapshots plus violation ageing mean improvement is slow and steady.

Crashes that were not your fault

Under FMCSA's Crash Preventability Determination Program, a crash of an eligible type can be submitted for review through DataQs as a Request for Data Review, with the police accident report and any photos or video. For crashes on or after December 1, 2024 there are 21 eligible types, including being struck in the rear, being hit while legally stopped or parked, and another driver going the wrong way or running a signal, plus any crash where video shows the sequence of events. Crashes older than five years cannot be reviewed. A crash found not preventable is removed from the Crash Indicator calculation, though it still appears on the record with the determination noted.

Changes that have been announced

In November 2024, FMCSA published an enhanced SMS methodology that renames the BASICs as compliance categories, folds Controlled Substances and Alcohol into Unsafe Driving and simplifies severity weights. It said the launch date would be announced in a separate Federal Register notice. As of this review, that notice has not appeared and FMCSA's methodology document, last revised in 2026, still describes the system set out on this page. The FAST Act limits on displaying property carriers' percentiles continue under the new methodology.

Common mistakes

• Assuming no percentile means a clean record. It may only mean too few inspections.
• Filing DataQs challenges without evidence. A request needs documents, not disagreement.
• Letting the MCS-150 go stale. Mileage and power unit counts feed parts of the calculation; see the biennial update.

Frequently asked questions

Is a high CSA score good or bad?

Bad. The scale runs 0 to 100 with 100 the worst performance.

Why did my score jump after one inspection?

Because small carriers have few safety events, so each one carries disproportionate weight. The same violation barely registers for a large fleet.

Can brokers see my CSA scores?

They can see your measures and the inspection and violation history in five BASICs: Unsafe Driving, Hours-of-Service, Driver Fitness, Controlled Substances and Alcohol, and Vehicle Maintenance. For property carriers, percentiles and alerts are not publicly displayed under the FAST Act. Hazardous Materials and Crash Indicator are hidden from public view entirely.

How often does the data update?

Monthly, from a snapshot of FMCSA's national database.

How long does a violation affect my score?

Up to 24 months. It carries a time weight of 3 for its first six months, 2 from six to twelve months, and 1 from twelve to twenty-four months, then drops out.

Can a crash that was not my fault be removed?

If it is one of the eligible crash types, you can submit it through DataQs under the Crash Preventability Determination Program. A crash found not preventable is removed from the Crash Indicator calculation.

Related

• Driver vehicle inspection reports.
• Hours of service explained.
• All compliance obligations.