using 2003. Anyone have experience with the PERT Analysis feature?
Is it useful? How? Can it be used in place of @Risk et al?
meg99
While PERT Analysis is still in wide use, it has been proven some years ago
to be statistically unsound. I'm not a statistician, but it general terms,
it accounts only for the probability of each individual task's uncertainty
and does not account for the accumulated uncertainty along a given path,
which is considerably greater than the individual task's uncertainty. IOW,
the resultant duration corrections are considerably smaller than they
actually should be had you done them using Monte Carlo analysis, which is
what the more sophisticated risk tools do.
My advice: Find a good risk analysis tool and forget PERT. I imagine others
will disagree, but if they do, ask them to cite a solid statistical
reference validating PERT.