Consider that one task leader might execute a task like his or her project task 100 times. Experience discloses that once or twice everything falls together, and that the task is completed in an unexpectedly short time. The same experience guides the task leader to recognize that many of the 100 tasks being undertaken will finish a little later, but not much later. There will be little variation in the finish times for a large number of the 100 tasks. This is because experienced people are doing what they know how to do, thus their outcomes are not going to vary.
There will be a large number of tasks within the hypothetical 100 that finish in about the same amount of time. There also will be some that take longer and, generally, one or two worst cases that take the longest. If our imaginary task leader records the finish time for each task, he or she can number the tasks according to their finish time. The shortest task is named number one. The longest task is named number 100. If the task leader plots the task durations on a vertical scale against the task number on a horizontal scale, the plot will look like the Figure 8-1.
There is a general pattern for task duration distributions for tasks that are repeatedly performed by knowledgeable and experienced people. In Figure 8-1, break points where task duration exceeds it and the magnitude of D, vary from task type to task type. (W represents the worst case duration.) For a routine task, the break point is likely to occur at Task 95 it is possible that one task leader will insist that this is a very familiar task and it can always be finished on time. For a familiar task that is not technically challenging but involves some judgmental decisions, the break point the point where the task's duration exceeds M may be expected to occur at Task 85. Fifteen out of 100 tasks can require some extra effort, and one or two will require an effort equal to W. For tasks involving some experimentation to get it right or some uncertainty about the availability of a key work-group member, the break point may occur at Task 75. For tasks that are particularly challenging or where resource availability is uncertain, the break point may be at Task 60. We expect experienced people to contribute an M estimate that is achievable at least 50 percent of the time (that is, for tasks 1 50). If this is not possible, the task leaders then are involved in a research effort, not a plannable project completion effort.