Who’s at fault · Rear-end collision

I Rear-Ended Someone Who Stopped Suddenly — Who Is at Fault?

AccidentLawyerReview Editorial Team · Updated 29 August 2026

A sudden stop does not by itself remove the following driver's duty to keep a stopping distance, so the presumption that starts there usually survives it. What has shifted the allocation is narrower: a stop made without cause, brake lamps that were not working, or a vehicle that merged in and braked immediately. Those are questions of fact, and the federal crash record rarely documents them.

AccidentLawyerReview is not a law firm. This is general information about how this situation is usually handled — not legal advice about a specific crash.

The diagram shows the fact pattern in its narrow form: two vehicles in one lane, the lead vehicle braking hard, and a following vehicle with less distance than the stop requires. That geometry does not change which vehicle carried the duty to keep a stopping distance. It changes what the record has to show, because several quite different collisions look like this one from behind, and they are told apart by the lead vehicle’s brake lamps, its lane position in the seconds before, and why it stopped.

What the sudden stop does and does not change

This describes how insurers and courts have generally approached this type of collision. It is not a prediction about any specific case, and fault is decided on the facts of each crash.

The five rebuttals are not variations on one argument: each turns on a different fact and a different record. The general presumption itself, and the way each state’s negligence rule converts a split into money, sit on the main rear-end fault page rather than here.

The four sub-cases

The stop had no cause. This is the only rebuttal that attacks the stop itself. It requires showing what was, or was not, in front of the lead vehicle, which means the record has to reach past the two vehicles that collided. The two codes cited above do not reach it identically: California prohibits bringing a vehicle to a complete stop so as to impede traffic, while Florida’s provision is written for slow speed rather than a stop. Both carve out a reduction that is necessary for safe operation or required by law — so establishing the absence of a cause is the whole of the work.

The brake lamps were not working. The most checkable of the five and the most perishable: settled at the roadside in a minute, close to unrecoverable a week later once the vehicle has been repaired or released. Statutes in both states require stop lamps, which is what makes an inoperative lamp a rule breach rather than a complaint.

The lead vehicle merged in, then braked. Here the argument is not that the stop was improper but that there was no interval in which a lawful following distance could exist. The evidence is positional rather than mechanical: lane markings, damage geometry across the rear quarter, and video. This sub-case is coded in the crash files as a lane-change event rather than a rear-end one, so it sits outside every figure below.

The stop was deliberate. Brake-checking differs in kind: the disputed fact is the lead driver’s purpose, not their conduct. The signalling and impeding statutes reach it, but purpose is not recoverable from damage geometry and the crash record rarely carries it, so this sub-case rises or falls on footage.

What separates them

Each sub-case is settled by a different record, and the half-lives of those records differ by orders of magnitude — one is gone within minutes of the vehicles being moved, another survives as long as the vehicle does.

The lead vehicle’s stop lamps

Decides the brake-lamp sub-case and nothing else. Checkable in seconds at the scene by the officer, the other driver or a passer-by; effectively unrecoverable once the vehicle is repaired, sold or released for salvage. It is the shortest-lived fact on the list and the easiest to capture.

Event data recorder, both vehicles

Most modern passenger vehicles hold several seconds of pre-crash speed, brake and throttle data. On the lead vehicle it shows how hard and how abruptly the stop was made; on the following vehicle it shows whether the brakes were applied at all, which is the question the crash report usually leaves open. Retrieval needs the physical vehicle.

Lane position before the stop

The only thing that separates a merge-then-brake from an ordinary sudden stop, and the two are indistinguishable from the damage alone once both vehicles are in one lane. Video, lane-marking scuffs and the position of the contact across the rear panel carry it.

What was ahead of the lead vehicle

The stop-without-cause sub-case is the only one whose evidence lies outside the two vehicles involved: a signal phase, a queue, a turning vehicle, a pedestrian. Witness accounts and traffic-camera coverage of the approach are what reach it, and municipal systems commonly overwrite on a cycle measured in days.

Forward-facing video

The only record that speaks to purpose, which makes it the one that matters for a deliberate stop. It is also the only single source that can resolve all four sub-cases at once, which is why its retention period usually decides what is arguable.

How often the vehicle ahead was actually stopping

Rear-end crashes are usually discussed as one event. In the 2023 CRSS file they are three, distinguished by what the struck vehicle was doing at impact.

Computed from the CRSS 2023 national file, 1,609,750 weighted crashes with a lead-vehicle state coded. Taken from ACC_TYPE and cross-checked against P_CRASH2, an independently coded variable, which gives 60.9% / 23.7% / 15.4% — a largest disagreement of 1.9 percentage points.

So the crash a searcher means by “they stopped suddenly” is 360,911 crashes in 2023 on the ACC_TYPE coding above — a fifth of the rear-end total, and a distinct minority of it. The percentages that follow are computed instead over the 387,230 striking vehicles the independent P_CRASH2 variable puts in the same category, which is the wider of the two counts. This crash also happens in a different place: 47.5% occurred away from any junction, against 34.5% of rear-end crashes generally.

The fatal file splits differently again. Across FARS 2018–2024, 15,801 fatal front-to-rear crashes carry a lead-vehicle state, and the decelerating share there is 15.1% — lower than in the all-severity file, where the largest fatal category is instead a lead vehicle moving at a lower steady speed, at 47.7%. The sudden stop is a smaller part of the fatal picture than of the everyday one.

What the crash record does not capture

The presumption is strong and the paperwork behind it is thin. These are the same 2023 crashes, counted by what the attending officer actually coded.

CRSS 2023. Following-driver figures are over 1,632,039 weighted vehicles whose critical pre-crash event was the vehicle ahead stopping, slowing or decelerating; lead-driver figures over 1,812,884.

Three of those rows bear directly on this fact pattern. A following-too-closely violation — the paper trace of the presumption — appears in 18.1% of these crashes and 18.8% of the decelerating-lead subset, so the citation record is not what carries the presumption. Whether the following driver braked at all is unrecorded in 72.7% of cases, which is the gap event-data-recorder readings fill and the reason a released vehicle is a closed question. And on the lead side the two circumstances the first and second sub-cases turn on are close to absent from the file: the driver-related factor for stopping in the roadway was coded on an estimated 7,745 vehicles (0.43%), and a lighting defect of any kind on 1,282 (0.07%).

The distraction rows point the same way. Distraction or inattention was affirmatively coded against the following driver in 17.6% of the decelerating-lead subset, while in 46.4% of rear-end crashes generally the field was not completed either way — a blank is an unanswered question rather than a finding of no distraction, and it is the single most common state of that field.

Absence of a code is not evidence that the circumstance was absent: CRSS records what was reported, not what happened, and its non-response rates are highest on exactly the fields this fact pattern turns on. What the file does describe honestly is the starting position. On the four sub-cases above the crash report is usually silent, and the material that is not silent is the material that expires first.

Keyword set verified against live search results on 2026-08-29.