The paper calls attention to the help which statistical technique can give to the commercial and engineering sides of electricity supply.A brief examination of the scope of the statistical approach shows that this technique is the key to many problems ranging from the action to be taken against hostile aircraft to the co-ordination of insulation. It enables experiments carried out “in the field” to be analysed, as it is possible with suitable planning of experiments to separate subsequently the various factors affecting a result.After a very condensed summary of the more generally used statistical methods, the application of some of these to generating stations and transmission systems is illustrated, particularly in connection with assessment of plant reliability.The present emphasis on planning to meet future loads makes forecasting a very important item. The Section dealing with this includes the effect of weather conditions on system load.The margin for random selection to be allowed in the setting of relays is considered; it is also shown how, in calculating the capacity of low-voltage lines, probability considerations directly affect the comparison between single- and multi-phase systems, and how statistical technique helps toclarify investigations into the voltage actually received at consumers' premises, and the reliability of various methods of earthing.It is suggested that rigid plus-and-minus limits may not always be the correct method of specifying restrictions on voltage drop. The argument of rigid limits is extended to meter testing, and by way of contrast the technique used in radio-interference testing is described.In dealing with matters affecting the consumer a prime consideration is the question of tariffs. Methods of gauging the effect of such changes in tariffs are indicated. The difficulty of ascertaining details about consumers' installations is greatly reduced by sampling. The use of sampling to enumerate details of the distribution system is referred to earlier in the paper.The importance of diversity, from the points of view both of tariffs and of system design, needs little stressing. Various methods of estimating diversity from knowledge of the consumer's load are discussed.Finally, one example of the replacement of wasting assets—street lamps—is analysed to show that planned replacement before the assets have worn out may be the most economical plan in certain circumstances.