If an analysis is made of the speed of a ship using data recorded in log-books the results are extremely disappointing, and do not reflect the accuracies of the instrumentation in use. Even data from a satellite navigator, where the speeds should have an accuracy of better than I per cent, give a I-knot band in recorded speed; in a 20-knot ship this is 5. per cent. Such inaccuracies require explanation.

To begin with one must ask what is meant by ship's speed, and different interpretations come from different people. The navigator is concerned with speed over the ground, while the naval architect or marine engineer is mainly interested in the forward speed of the ship through the water. On the other hand a compiler of ocean current data will require both the speed over the ground and the speed made good through the water, but the latter must take leeway into account. Finally, a weather routing consultant is looking for the actual forward speed the ship can achieve through the water in the prevailing weather conditions. All users of speed data may require other related information, such as the ocean currents and the leeway characteristics of the ship, which can in fact only be compiled from ship speed data. At the present time all these different requirements have to be satisfied by the speeds measured at sea and recorded in the logbooks and their abstracts. Ships' officers compile these records, but by the nature of their work they are biased towards the navigational concept of speed.