Risk and Decision Analysis in Projects, 3.0 Edition

Revisions and Errata

A revised 3.1 Edition is underway

Estimated publication in late-August, 2018
Short questions at chapter ends (for university classes using this book).
Other embellishments include: The "Critical Chain Simulation" video, enhanced Utility Elicitation Program , a regret table solution for VOI, and stochastic earned value analysis. The videos and updated UEP are available now.

Version 3.0 used “task” in most places in deference to Microsoft Project.
"Activity" is much prevalent in the project management literature (with tasks being a possible component).
Excepting Chapters 20 (MS Project) and 24 (CCPM modeleling), the new 3.1 edition generally uses “activity.”

Page 27, Table of EVM Variables and Formulas
The formula for the last item should be SAC = ET + STC

Page 113, 132, and 133. SEM confidence interval
The standard error of the mean, (SEM), describes how much error is in the EMV calculation and not the confidence range for the outcome NPV. More MCS trials improves the EV calculation precision (as 1/sqrt(no. trials)), but the NPV frequency distribution doesn’t change much.

Page 163, VII calculation. The $92k amount is correct, but the two labels are reversed.
Setting the inspection cost to $0 and no delay, EV NPV for Inspect Used Plant is -$4367k.
The Used Plant option without inspection EV NPV is -$4459k.
VII is the improvement, -$4367k − -$4459k = $92k.

Page 174,Project Risk Management, Project Budgeting, Differences
The Base Plan practice is unsettled, in my opinion. Here is my revised understanding of common practice:

   Base Plan is the sum of optimistic activity estimates

   Cost Estimate = Base Plan + Contingency
      where Contingency = sum of the risk EVs (from the Risk Register)

While Base Plan + Contingency may be a useful approach for budgeting and management, it is not my recommendation for estimation.

Instead, a full stochastic model does away with contingency. The analysis aim is an unbiased estimate using the EV concept.

Page 328, Project Progress: The stochastic earned value chart shows a confidence envelope extending the Planned Value (PV) Curve. This is incorrect. PV does not change unless there is an approved scope change.

Page 366, One-at-a-Time (variance analysis): This method DOES NOT show or include stochastic variance. This method itemizes individual variable deterministic variance. Compare the total of these to the NPV when all variables are set to their actuals (or revised assessments). That difference is the total of joint variances.

Please contact me if you have questions or concerns about anything you have read in the book. E-mail:

Press your browser's <back> key to return.