23 April 2012

Why Critical Path Scheduling (CPM) is Wildly Optimistic

Pat Weaver has a really well written paper about the fatal problems with CPM. He goes beyond the flaw of averages issues and covers the topic in depth.

I'm not sure I agree with all his prescriptions, but it does have me thinking about my preconceptions.

The abstract and link to download the pdf are here.


  1. An approach that I've found to be helpful is to develop a tornado analysis of the activity durations in the schedule on the end date in addition to the traditional critical path. Based on the simulation of the schedule, then, you get a rank ordered list of activity durations that contribute the greatest likelihood of creating schedule delays. If the task durations are calibrated properly, you then have information about how to manage the causes of variance that lead to schedule overruns.

  2. Hi Rob,

    You're still calculating with point values in an environment that has Jensen's inequality affecting just about every activity and warping your results. That's what we're trying to fix with Probability Management.

    One of the advantages is that sensitivity analysis takes just a scatterplot of the activity distribution against the final milestone distribution. Very visual and easy to explain.

  3. No, my algorithm for tornado analysis actually adjusts for Jensen's. I didn't know it at the time I developed it in 2000. I just recognized that the traditional approach to tornado analysis wasn't adjusting for nonlinearities in the objective function. It was after I read Savage's work that I realized that I had already taken Jensen's into account. My value of information algorithm adjusts similarly.