A closer is not a second incident

If you add a stay-away line to the walk-off, the no-show, and the missing fuel card, you will count the same review twice and call the pile a risk score. The closer is the last carrier talking to you. The scene is what happened. Those are different registers. Treat them as one number and the file gets louder without getting clearer.

Among the reviews carriers left on CDLScan, 32.2% contain a verdict. That figure describes those reviews, not drivers in general: carriers write after a bad run, so the share tells you what shows up once someone bothers to report it. Abandoned-truck language sits in 33.5% of the same corpus. Those two shares are not two populations. Stack them and you double-count write-ups that already contain both a scene and a closer.

Two registers, one paragraph

Incident language is physical. Abandoned truck, left the truck, never showed, flight ticket, stopped answering, cash advance, damaged trailer. You can point at a unit, a day, a load, a phone.

Verdict language is a recommendation to the next desk. Stay away, waste your time, don’t hire, avoid this driver. You cannot point at a unit. You can only hear how done the author is.

The phrasing split is the whole method. If you read stay away as if it were left the truck, you have turned heat into an event. If you ignore left the truck because the closer is missing, you have turned a missing recommendation into a pass.

What rides with a verdict looks like the corpus itself. An abandoned truck sits in 32.3% of verdict reviews — against a 33.5% base rate, essentially chance. Aggression sits in 16.0%. Drugs or alcohol sit in 14.3%. A crash or damage sits in 13.9%. Late or missed deliveries sit in 11.1%. Theft sits in 10.3%. Going unreachable sits in 9.5%. Paid travel with no arrival sits in 7.9%.

That mix is the proof you needed. A stay-away line does not point at one failure type. It points at a carrier who decided to warn you. The useful screen is still the scene.

How we count these shares, and why year totals cannot be compared, is on the methodology page. The companion on the closer as heat is what do-not-hire wording actually tells you. The floor under almost every write-up is in what a lowest rating actually contains. Why the corpus is already a complaint file is in carrier review bias.

Why desks double-count

Because both lines feel like findings. Because a 1-star write-up already feels like a verdict, and 94.8% of this corpus already sits on that floor. Because a desperate recruiter would rather tally strikes than read the sentence that names the truck.

A stay-away closer on an abandoned-truck review is not two strikes. It is one walk-off, written with heat. A stay-away closer on a no-show is not two strikes. It is one empty chair, written with heat. A stay-away closer with no scene is heat without an event — still a reason to look the name up, not a reason to invent a second category.

The other error is the mirror image: treating a review with no closer as clean. Most write-ups do not contain verdict language. 32.2% do, which means the rest do not. An abandoned truck does not need stay away sitting next to it to be a walk-off screen. A no-show does not need don’t hire to be an empty chair. The event does the work. The closer is optional.

The closer itself has climbed. Verdict language ran 19.6% of reviews in 2018 and 37.2% in 2025. I stop that line at 2025. Carriers are ending more write-ups with a recommendation. That is a fact about how reviews are written, not a fact about how drivers behave, and it is not a second incident rate.

What official reports do with a closer

Nothing. PSP, an MVR, and the Clearinghouse do not store stay away. They store filed crashes, inspections, license actions, program events. A review that ends in don’t hire and a review that ends with a date and a location look the same to those systems: they look like nothing.

So a clean packet on a stay-away history is the default, not a finding. The filing systems were not built to hold a carrier’s recommendation to the next desk. The review is. Read the event. Use the verdict only as heat. Do not add the heat to the event and call the sum a score.

I am not going to pretend this is an independent study. CDLScan is the source, and I write from the data they share. Verdict language is in about a third of what carriers bother to write down. Incident language is in the sentences that came first. Adding those shares is the fastest way to misread the file.

What was visible before you spent

The tell is the gap between the last line and the rest of the paragraph. Typical sequence, built from the mix, not from one person’s file: you open a profile, see stay away, write “do not hire” in your notes, and stop. Or you see no closer and skip the write-up. In both cases you skipped the sentence that would have changed the hire.

If the sentence is left the truck, dropped the truck, abandoned the equipment — treat it as a walk-off screen even if the closer is loud. Ask where the last truck was left, who recovered it, and whether the trailer was still loaded.

If the sentence is never showed, flight ticket, ticket never — treat it as an onboarding screen. The closer does not turn an empty chair into a walk-off.

If the sentence is stay away and nothing else, you have heat without a scene. Look the name up. Do not copy the closer into your own notes as if you had verified an event.

Do not average closers and call it a score. 32.2% of this corpus already ends in a recommendation. An average here measures how often carriers warn the next desk, not how often a driver fails.

Then look the name up across write-ups, not across last lines. Check it on CDLScan before you spend. It takes less than a minute. The useful part is the scene, not a running total of stay-away lines.

Frequently asked questions

Can I add the verdict share to the abandonment share?

No. They are different registers. Many write-ups contain both a scene and a closer. Stacking 32.2% and 33.5% double-counts the same review.

Does 32.2% mean a third of drivers should not be hired?

No. 32.2% is the share of reviews, among those carriers left on CDLScan, that contain a verdict. It is the author’s recommendation to the next desk, not a fact about the driver, and it describes those reviews, not drivers in general.

If there is no stay-away line, is the file clean?

No. An abandoned truck, a no-show, or a missing fuel card does the work without a closer. Read the scene. Do not treat a missing recommendation as a pass.

Is stay away a behavior I can screen for?

No. Stay away is attitude toward the reader. Left the truck is the screen. Search the event language. Use the closer only to know how done the last carrier was.

Will a clean PSP tell me the last carrier overreacted?

No. Official reports do not store stay away. A clean packet means the filed items are clean. It does not classify a closer, and it does not explain a scene.

What should I tell a hiring manager who quotes the last line?

Give them the event and the year, not the closer. Stay away is heat. The sentence before it is the hire.

Where do I read the write-ups?

On CDLScan, before you hire. Look at what happened. Do not add the closer to the event and call the pile a score.