Can you tell how fast a motorcycle was going in an accident?

Yes, by three routes: the rotation it imparted to the struck vehicle, the reduction in the wheelbase of the bike, and the engine revs with the gear engaged. Each has its own range of application.

Which is the most reliable method for motorcycle speed?

The one based on the rotation of the struck vehicle, validated by its authors against twelve instrumented tests with R² = 0.9925 and a mean error of −0.81%. It requires the struck vehicle to have rotated measurably.

Why does fork deformation underestimate the speed?

Because there comes a point where the fork and wheel are already crushed against the frame and the engine, and the wheelbase stops shortening. That saturation is around 52 km/h, and above it the result should be treated as a lower bound.

Does the same method as for a car apply?

No. The CRASH3 method relies on stiffness coefficients from barrier tests on cars, and a motorcycle structure does not respond in the same way. Specific correlations and models are used instead.

How to calculate the impact speed of a motorcycle

Updated 2026 · 5 min read

The motorcycle is the hardest case in a reconstruction: it leaves almost no skid mark, it goes down on its side, and its deformation cannot be treated like a car's. There are three routes, and none of them serves every case.

1 · From the rotation imparted to the struck vehicle

When a motorcycle strikes a car away from its centre of mass, it makes it rotate. That rotation is measurable —it is recorded in the marks and in the rest position— and the motorcycle speed can be derived from it, because the angular momentum the car received had to come from somewhere.

This is the best supported of the three: its authors validated it against twelve instrumented tests with a coefficient of determination of 0.9925, a mean error of −0.81% and no point outside ±10%.

When it cannot be used

It requires the struck vehicle to have rotated measurably. If the impact passed almost through its centre of mass, the lever arm is very small, hardly any rotation is generated and the method loses sensitivity: the result swings wildly with any measurement error. In that situation another route is needed.

2 · From the wheelbase reduction of the motorcycle

In a frontal impact the fork compresses and the distance between axles shortens. That reduction is measured with a tape and correlated with speed. It is the practical route when the other vehicle did not rotate.

It carries a limitation that must always be declared:

The method underestimates above a certain speed

Once the fork and the wheel have been crushed against the frame and the engine, the wheelbase stops shortening no matter how much speed there is. That saturation is placed at around 52 km/h for the class of motorcycle tested. Above that the result is a lower bound: the real speed may have been considerably higher.

The correlation also needs the crush of the struck vehicle. Correlating with the wheelbase alone is an older model whose own authors measure a correlation of 0.27 and discard as unreliable.

3 · From engine revs and gear

If the gear engaged and the revolutions at the moment of impact are known —from a jammed tachometer, from the ECU or from the mechanicals themselves— the speed follows from the transmission ratio and the rolling radius. It is exact when the datum is reliable, and worth nothing when it is not.

What to do with all three

Compute with every one the data allow and see whether they agree. Each fails in a different place: the rotation method needs rotation, the wheelbase method saturates at the top, the engine method depends on an external datum. When two agree, the result stands; when they do not, the discrepancy is itself information about which input is wrong.

Module 06 brings together the three methods, estimates the yaw inertia of the struck vehicle when it is not known, and warns of each limit of application.

Open the calculator →

Frequently asked questions

Can you tell how fast a motorcycle was going in an accident?: Yes, by three routes: the rotation it imparted to the struck vehicle, the reduction in the wheelbase of the bike, and the engine revs with the gear engaged. Each has its own range of application.
Which is the most reliable method for motorcycle speed?: The one based on the rotation of the struck vehicle, validated by its authors against twelve instrumented tests with R² = 0.9925 and a mean error of −0.81%. It requires the struck vehicle to have rotated measurably.
Why does fork deformation underestimate the speed?: Because there comes a point where the fork and wheel are already crushed against the frame and the engine, and the wheelbase stops shortening. That saturation is around 52 km/h, and above it the result should be treated as a lower bound.
Does the same method as for a car apply?: No. The CRASH3 method relies on stiffness coefficients from barrier tests on cars, and a motorcycle structure does not respond in the same way. Specific correlations and models are used instead.

Sources

On the limits of the method when applied to a motorcycle:

McHenry, R. R. (1976). User’s Manual for the CRASH Computer Program. Calspan ZQ-5708-V-3. Its Exhibit 9-5 lists fourteen surfaces with four pairs of friction values each. The CRASH3 stiffness coefficients come from barrier tests on passenger cars.

How these methods compare against documented cases is detailed in How this calculator is validated.