The problem of collision avoidance at sea would be an important one even if accidents were rare, but the evidence suggests that they are much more common than one might expect. In the eight years leading up to May 1963 there were 60 collisions in the English Channel area and in the following eight years the number increased to 94. With the development of navigational aids such as radar one might have hoped for a decrease. The suggestion that the increase is quite likely attributable to the possession of radar is ironic indeed. It would appear that having the use of this ‘safety’ equipment allows the mariner to take greater risks, but that the risks which are in fact taken are greater than the equipment allows. We would do better to go back to fog horns alone. Even if over-confidence is not the only factor responsible for the increase in accidents, it does, nevertheless, illustrate the importance of the psychological variable in any man–machine system.Collisions at sea, however, are not simply due to errors in a man–machine system, It is, of course, the hardware that actually collides, but the cause of the interaction lies in the decisions and indecisions of the two human agents controlling the vessels. It is essentially a social encounter. It is for this reason that the work initiated by Captain Kemp is greatly to be welcomed.