The present paper attempts at a contribution to peak load pricing, in both theory and application. The general result from the traditional theory that charges the off-peak consumers marginal operating costs only and the peak users marginal operating plus marginal capacity costs, since it is the on-peakers who press against capacity, has already been called into question in the literature. It has also been shown that the equity norms are violated in the traditional peak load pricing, whereby off-peak users pay no capacity charges, but are supplied output out of the capacity, `bought/hired' by the on-peakers. Theoretical attempts at modification have proved that the traditional conclusion holds only for homogeneous plant capacity (e.g., in one plant case), and in economic loading of two or more plants, the off-peak price also includes a part of capacity costs. This paper, however, shows that if the off-peak period output is explicitly expressed in terms of capacity utilisation of that period, the result will be an off-peak price including a fraction of the capacity cost in proportion to its significance relative to total utilisation. This would appear as a general case, irrespective of the nature of generation technology, that is, even when there is only one plant. We also give an illustration by estimating marginal costs and peak load prices using time series data on the Kerala power system. Where the data are incapable of yielding the required statistically determined long-run relationship among the variables under study, we propose a simple and viable method of using discrete ratio of increments in lieu of a marginal value.