An alternative to estimating a continuous concentration curve is to evaluate the integral in equation (2) directly as a weighted sum of the observed instantaneous loads:
where
is the instantaneous load at 
.
If load is measured uniformly in time, or at random, the same weight might
be applied to each instantaneous load. In this case the value of 
would depend only on the number of measured concentrations per unit time.
However, more precise load estimates can usually be obtained by selectively
sampling at those times when uncertainty about the instantaneous load is
greatest. Under these circumstances, the weights will vary.
Direct methods have some important advantages over the rating curve. In general they require fewer assumptions about the underlying physical process. Also, because direct methods do not involve transformations of the data, they are not subject to retransformation bias.