Your Unsafe Driving BASIC Is Clean. Your Crash Indicator Is 92nd Percentile. Guess Which One the Plaintiff Shows the Jury.

A broker I know ran a carrier file last fall. I'll call the carrier MC-1847293 / DOT-4059317. Eleven months of authority — right at the edge of the eighteen-month caution zone. Forty-two power units, mostly dry van. He pulled the SAFER snapshot. Checked all the boxes: Unsafe Driving BASIC below the alert threshold, OOS rate at 8.4%, cargo policy on file with a solid insurer, $750K BIPD confirmed through FMCSA's L&I.

He booked them. Not once — he ran them for three months on a dense retail distribution lane, thirty-plus loads, home goods and furniture averaging $50,000 to $70,000 per trailer.

Then one of their drivers hit a passenger vehicle at a highway merge on I-64. Two fatalities. By the time the insurer and the plaintiff's attorneys were done mapping the broker's carrier file, the broker was named in the suit. And the first slide the plaintiff's attorney showed in the deposition wasn't the driver's PSP record. It wasn't the OOS rate. It was a single number.

Crash Indicator: 92nd percentile.

"You ran your Unsafe Driving check," the attorney said. "Did you look at this one?"

My broker friend didn't have a good answer.

What these two BASICs actually measure

The FMCSA Safety Measurement System has seven BASIC categories. Two of them specifically relate to collision risk in ways that routinely trip up brokers: Unsafe Driving and Crash Indicator. Most brokers I talk to check Unsafe Driving first. Some treat Crash Indicator as a secondary glance — if it's below the alert threshold, fine, move on.

That's the wrong read.

Unsafe Driving measures roadside inspection violations — speeding, improper lane changes, texting, following distance, reckless driving citations. It's forward-looking behavior as captured by FMCSA inspections. The problem: a carrier that drives dangerously but rarely gets pulled for inspection won't surface there. You're seeing what inspectors caught, not what the carrier actually does between inspections.

Crash Indicator is different. It measures crash history from the FMCSA Motor Carrier Crash database — crashes reported by law enforcement, weighted by severity. A fatal crash counts more than a tow-away. A tow-away counts more than property damage only. FMCSA publishes these figures normalized per power unit and scales them into a percentile against all carriers nationally. You're not looking at predicted behavior. You're looking at outcomes that already happened.

FMCSA uses the 85th percentile as the intervention threshold. Carriers above it get an alert flag in the SMS and move up the queue for compliance reviews. If FMCSA has decided this carrier's crash history warrants closer attention — and you selected them anyway without noting why — that's a question you'll eventually answer in a deposition.

Why the Crash Indicator hits different in court

A high Unsafe Driving BASIC is bad. But a plaintiff's attorney is going to work to explain it away on your behalf and then pivot. They'll say the BASIC shows the carrier was likely to drive unsafely. That's probabilistic. Juries don't always track probabilistic reasoning well, and defense attorneys can muddy it.

Crash Indicator doesn't require a prediction. It says crashes happened. Law enforcement filed reports with FMCSA. The carrier's trucks have been involved in serious incidents at a rate that puts them in the top fifteen percent of carriers nationwide.

That sentence plays to a jury. It's not a model. It's not a regression estimate. It's a documented record of outcomes the federal government compiled from police reports.

And here's the other problem: crashes reported to FMCSA don't require an inspection trigger — any crash meeting the severity threshold gets reported by the investigating officer. That means Crash Indicator doesn't depend on whether a carrier is routinely pulled over. A carrier that manages to avoid roadside inspections can still accumulate crashes. Those crashes will appear in the Crash Indicator even if Unsafe Driving is clean.

A clean Unsafe Driving BASIC on a carrier with an elevated Crash Indicator isn't reassurance. It might mean the carrier is good at avoiding inspectors and bad at avoiding crashes.

What the plaintiff's attorney is looking for

Post-Montgomery v. Caribe Transport II, when a plaintiff's attorney is building a negligent selection case against a broker, the first FMCSA data pull on your carrier file is usually the SMS BASIC printout. They're looking for Crash Indicator percentile.

Here's why: it's defensible for them. The data comes from FMCSA. It's based on law enforcement reports. The defendant broker can't argue the data is biased or incomplete — FMCSA published it, it's publicly available on SAFER, anyone with thirty seconds and an internet connection can pull it. If you didn't look at it: you failed to check publicly available government data showing this carrier had a documented crash history. If you did look at it: you saw this number and decided to put drivers from this fleet on public highways anyway. What was your reasoning?

Neither position is comfortable if you don't have documentation.

Now add the actual numbers. MC-1847293 at 42 power units, Crash Indicator at 92nd percentile. At that fleet size, average carriers show somewhere in the range of zero to three FMCSA-reportable crashes over a two-year window. A carrier at the 92nd percentile — depending on severity weighting — is likely showing six or more. That's roughly double the industry median crash rate, normalized per power unit.

The plaintiff's attorney is going to run that math in their opening statement. Twice the industry average crash rate. You booked thirty loads.

The FMCSR context

49 CFR § 385.5 defines the safety fitness standard carriers must demonstrate to stay operating. The statute requires that carriers maintain safety management controls adequate to ensure compliance with applicable regulations. FMCSA's SMS — including the Crash Indicator BASIC — is how they operationalize that standard to decide which carriers need intervention.

When a carrier's Crash Indicator puts them above the alert threshold, FMCSA has flagged them for closer attention. That's not a legal disqualification — you can still book a carrier in alert status, and many of them are running fine despite the flag. But the flag exists. A federal agency put it there. If you selected them and something goes wrong, the jury learns that the government had already identified this carrier as one worth watching.

That's not an insurmountable fact. It is a fact you need to be prepared to address.

What to do when these two BASICs conflict

The right move isn't to refuse every carrier with an elevated Crash Indicator. It's to dig deeper when they conflict — clean Unsafe Driving, elevated Crash Indicator — and document your reasoning if you proceed.

When I see that combination, I want answers to specific questions before I make a decision.

How many crashes, and what severity? FMCSA breaks out the crash count by fatal, injury, and tow-away. A carrier with three tow-away crashes in two years on a 40-unit fleet is different from a carrier with one fatal crash. Same high Crash Indicator percentile, very different risk profile.

Is the trajectory improving or getting worse? The Crash Indicator runs on a rolling window. If the carrier had a rough period eighteen months ago and the number has been declining since — that matters. If it's been climbing quarter over quarter, that matters more.

What does the crash preventability database say? FMCSA's Request for Data (RDR) program lets carriers challenge crash preventability determinations. If a carrier has successfully had crashes marked "not preventable" — the other vehicle crossed the centerline, a ladder fell off a truck ahead of them — those crashes carry different weight. The preventability data doesn't erase the Crash Indicator, but it adds context that belongs in your reasoning.

Can you explain your decision in plain English to a jury? That's the test. If you looked at a 92nd-percentile Crash Indicator, understood what it meant, noted the severity breakdown, saw an improving trajectory, and made a documented choice to proceed on a defined lane for a defined set of loads — that's defensible. If you saw the number and kept moving because Unsafe Driving was clean, that's the answer that ends careers.

What I actually do with conflicting BASICs

When DOTScreener returns a carrier screen and Crash Indicator is in the alert zone, I don't mark it as a pass and move on. I treat it as a file that requires a specific written note — not just a screenshot of the BASIC printout, but a human-readable observation.

The note covers: Crash Indicator percentile as of the screening date, number and severity of crashes in the reporting window, preventability determinations if available, whether the trajectory has been improving or worsening, and explicit reasoning if we're proceeding with the carrier.

The timestamped Carrier Selection Record in DOTScreener captures the BASIC snapshot the day I ran the screen. That becomes the documented record — it establishes what the data showed at the moment of selection, not what FMCSA shows six months later after something went wrong. And if the Crash Indicator moves after we're already running loads — say it crosses into alert territory three months in — Continuous Monitoring fires an alert, the file gets a new entry, and I have to make the same documented decision again.

That's the loop that builds a defensible carrier file. Not just checking boxes at onboarding, but tracking what changes and noting your response.

The one thing most brokers get wrong

Brokers treat Unsafe Driving as the primary crash-risk signal because it's intuitive. Drivers who speed get in crashes. Drivers who text get in crashes. The logic tracks.

But the Crash Indicator is telling you something Unsafe Driving can't: whether this carrier's fleet has actually produced crashes, regardless of what inspectors have caught. It's the outcome measure. Unsafe Driving is the leading indicator. Crash Indicator is the trailing reality.

When they conflict — driving behavior looks fine, crash outcomes look bad — the trailing reality should be making you ask harder questions. Not necessarily killing the carrier, but making you work to understand why the gap exists and whether you can explain it if you're ever asked.

A clean Unsafe Driving BASIC is not a reason to stop reading the screen. The Crash Indicator is still there. So is the question you'll eventually have to answer about whether you saw it.

How I document this

When the Crash Indicator is elevated, my file note looks like this:

"Carrier MC-1847293 screened [date]. Crash Indicator: 92nd percentile as of screening date. Crash breakdown: 2 injury, 4 tow-away in current 24-month window. No fatal crashes. No preventability challenges on record. Unsafe Driving: below alert threshold. Decision to proceed on [lane description] loads for [time period] based on: non-fatal crash profile, improving trajectory over last 6 months, [insert other factors]. Will re-evaluate if Crash Indicator moves or at 90-day interval."

That note lives in the file. It's dated. It shows I saw the number. It shows I understood what it meant. And it shows I made a deliberate, documented choice rather than a reflexive one.

If I'm wrong, the file still shows I wasn't negligent. That's the goal.

— Mason Lavallet

Founder, DOTScreener.com