The evaluation of the safety performance of passenger vehicles is in most test procedures restricted to the inherent safety. But in real-world accidents, the vehicle influences both its own passengers and the passengers of the colliding vehicle or the cyclists or the peclestrians. So test procedures that can only measure the performance of the vehicle and neglect the influence of the colliding vehicle will not be able to describe the safety performance completely.
Another problem is the separate evaluation of front-impact and side-impact. When side-impact protection is increased on the one hand side and the front of the vehicle becomes more stiff, due to increased front-impact protection, a high increase of safety will be measured, although no progress to side-impact protection may occur in real-world accidents.
By computer simulation, these effects can be described. A frontal impact with a movable deformable barrier only makes sense when the acceleration of the barrier is measured as well, and by mathematical simulation transformed to a dummy load of the "passenger of the struck vehicle". A side impact of the vehicle against a vehicle of the same type may show the "self-compatibility" and detect too stiff front-structures.