Best Practices for Buffers

[Gary Furash]

I was hoping to get any best practices for the creation
of Buffers (Contingency, etc.). That is, if a given task
or summary task (I have both situations):

- is likely to take X;
- but could take much more if something goes badly wrong,
or some possible, accepted risk occurs; and
- I have some idea how much extra it could take.

Things I've done before:
- Extended the expected duration of the task with notes
that it could take less.
- Created an associated buffer task, clearly identified
as such.

 

[Steve House]

Ed's advice is spot on and I'd concur. Never been happy with the idea of
buffer tasks because they seem to me to be introducing pseudo-events into
the plan purely to serve as fudge factors. By definition they distort the
schedule. PERT seems to me to be much more accurate - "This is what we
expect to be able to do but in the worst case it might go to this" showing
two different schedules.