I've included one new piece of information in Figure 9-21. Notice that column K is the average type (Avg Type). Two different types of average calculations are shown in Figure 9-21: straight and weighted.

While the straight average for the IO read wait times is not shown in the figure, it is calculated as follows:

However, the scattered read wait time of 0.093 ms occurs much more frequently than the other waits. A more accurate average calculation would take into account the occurrences of the scattered read waits in addition to its average event wait time. This is called a weighted average, and it is a much better average calculation when working with diverse and highly variant data sets, as we have here.