A pedestrian struck by a vehicle is a small mass hit by a large one, and that mismatch is what makes the event reconstructible. The body carries away a share of the striking vehicle's momentum, follows a path set largely by front-end geometry, and comes to rest somewhere measurable. Working backwards from that rest position to an impact speed is standard practice with a deep base of full-scale testing behind it. What it delivers, though, is a bounded range whose width depends on assumptions that have to be stated rather than buried.
Trajectory type is the first determination
Throw-distance analysis is not one equation. It separates by how the pedestrian and the vehicle front end interact: forward projection, where a high or blunt front strikes above the centre of mass and drives the body ahead of the vehicle; wrap, where a lower hood edge catches below it and the torso rotates onto the hood before sliding off; and the vault cases, where the body clears the fender or the roof.
Choosing the wrong family shifts the answer systematically rather than adding scatter. Front-end geometry, pedestrian stature and stance, and the height of contact damage discriminate between them, which is why the determination is made from physical evidence before any arithmetic begins.
What the models are actually fitted to
The commonly used relationships are empirical fits drawn from staged impacts, instrumented surrogate testing and reconstructed real-world cases accumulated over decades. That provenance is a genuine strength and also a limit: a fit describes the population it came from, and where the vehicle, pedestrian or surface sits at the edge of that population, the scatter band widens.
Contemporary front ends are not the front ends behind the earlier data sets. Hood height, bumper profile and energy-absorbing structure all change how a body is launched.
The inputs that move the answer
Sensitivity is where these opinions are won or lost. The effective friction the body met while sliding or tumbling is usually the largest single lever, and it is rarely one value: a body may cross asphalt, then a painted marking, then a gravel shoulder, each resisting differently. Pedestrian pre-impact motion matters too, since someone walking into the path carries lateral velocity the simple models ignore.
The remaining inputs — mass and stature, the vertical location of first contact, and the braking state at and after impact — each shift the result by a knowable amount. Quantifying that shift separates an estimate from an assertion.
Braking changes the geometry, not just the speed
A vehicle braking hard at impact pitches forward, lowering the bumper and hood edge relative to the pedestrian, and that alone can convert what would have been a wrap into a forward projection. Braking afterwards governs whether the body is carried or separates immediately. The braking question therefore feeds the trajectory determination rather than only the arithmetic, and it is where recorded vehicle data changes the analysis instead of merely confirming it.
Rest position is not the impact point
The model needs throw distance measured from the point of impact, and that point is not directly observable. It is inferred from debris origin, from where scuff or gouge marks begin, from fluid and glass distribution, and from the pedestrian's path into the roadway. Every uncertainty in that inference propagates straight into the result.
Post-impact movement compounds it. Bodies are dragged, roll, or are moved by responders before anyone documents the scene, and a rest position recorded after care has begun is not the measurement the model assumes.
Why a range is the correct output
Presenting a single figure from a throw-distance calculation misstates what the method can do. The defensible product is a range with its basis identified: which trajectory family, what friction band, what contact height, and what the result becomes when each input moves to the other end of its interval.
Most pedestrian matters do not turn on a few miles per hour either way. They turn on whether the vehicle was above or below a posted limit, or above or below the speed at which a stop was achievable in the distance available.
Independent evidence tightens the band
Throw distance is rarely used alone because other records constrain the same variable by different routes: recorded pre-crash speed, the height and lateral position of contact damage, injury contact points mapped to vehicle geometry, and frame-timed video.
Where two independent methods converge, the reconstruction is strong and any remaining disagreement is about inputs rather than physics. Where they diverge, that is itself a finding worth chasing, and it usually points at a mistaken assumption about pedestrian position.
Cases where the model does not apply at all
Low-speed backing incidents produce no meaningful throw. The pedestrian is knocked down and often contacted in place, and the determinative facts are the rear blind zone, mirror coverage, and whether a rear visibility system was fitted and working — the subject FMVSS 111 addresses for light vehicles.
The same caution applies to pedestrians struck while already prone, to contacts by a mirror or trailer rather than the front structure, and to multiple-contact sequences.
Where these opinions are challenged
Predictably: that the trajectory family was assumed rather than established from damage and injury geometry; that a single friction value stood in for a varied surface; that the impact point was placed where the conclusion needed it; that the rest position came from a diagram prepared after the scene was disturbed.
Work expressed as a range, with each input identified by source and each sensitivity quantified, absorbs that scrutiny. The SAE recommended practices covering reconstruction and recorder retrieval exist to make that traceability routine.
This article is general technical orientation, not a failure analysis, an engineering opinion, or advice on any specific matter. Determining the cause of a particular incident requires hands-on examination by a credentialed expert.