Blogs

Schedule Quantitative Risk Analysis: Duration Ranging

By Rufran C Frago posted 01-09-2021 06:39 AM

  

Everyone in the risk management community knows what schedule quantitative risk analysis is. It is what most of us in project management call the traditional method of quantification. One day, I decided to ask a group of subject matter experts about this.

"What do you think is the best approach to duration ranging?"

I have done and facilitated so many three-point activity duration estimating sessions that I have already lost count. Regardless of how many, though, I still ask myself, "Am I doing the three-point estimate correctly?"

Professional Opinion

There is a growing opinion from some risk management professionals that the "duration range" should only consider the estimation of uncertainties, not risk events. Let us take the activity called "piling installation." It is an activity with 20 days duration. It will probably take 20 days, but it can be 25 days at maximum and 18 days at a minimum.

This three-point spread here is because of estimating uncertainty. Risk event (or events) do not affect activity duration. We will consider risk events using the risk register (risk scoring).

Duration range must stick within the breadth and averages of the work itself based on normal conditions rather than assumptions of abnormal conditions.

Relative Values between Three Points

How about duration ranging? How should we monitor the validity of the three-point estimate? Should the risk analyst allow the following flawed values?

1) Most likely duration equals minimum duration; i.e., ML =MIN

2) Most likely duration equals maximum duration; i.e., ML=MAX

3) Most Likely Duration is less than Minimum Duration; i.e., ML < MIN

4) Most Likely is greater than Maximum Duration; i.e., ML > MAX

5) Most Likely is equal to Minimum and Maximum Duration; i.e., ML=MIN=MAX

6) Most Likely is different from the Remaining Duration; i.e., ML NE RD

These relative values are quite common when facilitating a risk session or running the risk analysis itself. The Default Most Likely (ML) value of each activity equals the activity's remaining duration.

The basic principle of schedule development calls for the schedule to represent the "most likely."

A BRIEF REFLECTION

In many cases, projects end up giving any of the numerous flawed inputs stated above. When this happens, the activity on focus is not a P50, or the Most Likely. When ML is not equal to RD, depending on the differential value and the frequency, the schedule says whether it is aggressive or lax. Once all the values are in, the risk analysis tool runs the simulation. It will calculate the inputs versus the network backdrop, providing the result or generate either a warning or an error. The errors and warnings usually boil down to preceding items 1 to 6.

A subject matter resource that inputs a value that has ML=MIN is trying to say that the remaining duration is flawed and does not represent the expected value of P50. When he brings the minimum value to equal the "most likely" value, he says that finishing the activity on time using the remaining duration has less probability.

It will be a more significant challenge than when the ML stays as is, and the minimum value opens to the left of any specific duration range. He thinks that the remaining duration of the activity is the earliest possible duration. The "remaining duration" of the task is nowhere near the most likely deterministic duration value.

Conversely, when ML=MAX, there is no way the activity will complete later than the activity's maximum remaining duration. We can also interpret the "Most Likely" value as too pessimistic that needs recalibration to a lesser value. The facilitator must clarify from the subject matter experts (SME) the rationale and act accordingly. By saying so, the project is assuring that the deterministic value is the same as the maximum.

Does this make sense? Something is not quite right if one thinks about it for a moment. If two of the three values were equal, would it still be a 3-point range? What is the right perspective? Are we going to allow this? If yes, why should we let this?

Many risk analysts still consider the right triangle distribution as the only believable distribution. Fortunately, the SQRA tool can successfully do the calculation. In other words, there is technically no problem. The facilitator should avoid such a distribution scenario because it is not a fair reflection of the possible ranges and can be self-defeating. Such distribution is unrealistic and better deemed impossible.

Items 3 and 4 will come out as an error. The calculation will stop. The Most Likely value lying outside the confines of minimum and maximum values is unacceptable. The process will not proceed unless we fix the issue. Calculation stops because value relationships do not make sense.

Again, when the Most Likely duration equals the Minimum and Maximum or ML=MIN=MAX, the warning message comes up. This kind of entry is the same as no entry at all if Remaining Duration is equal to Most Likely or RD=ML. Now, I'm sure you won't forget this simple rule. Since there is no duration range provided, it has practically no input to provide. You can ignore the warnings.

TRIGEN

A distribution rising in popularity is called the TRIGEN (P10/P90). The distribution profile where the minimum represents, for example, a P10 and your maximum a P90. The calculation considers what risk practitioners call "outliers," values that describe the remaining 10% outside the P10/P90 points.

When a subject matter expert provides duration input ranges based on their experience, the assumption considered is that their minimum and maximum value inputs are not accurate reflections of the actual minimum and maximum. 

Trigen Rationale

Even the experts have not experienced the real minimum and maximum values in their lifetime. To compensate for the blind spot, we opt for P10/P90 to consider outliers.

About the Author

Rufran C. Frago is the Founder of PM Solution Pro, a Calgary consultingproduct, and training services firm focusing on project and business management solutions. He is passionate providing advice, mentorship, education and training through consultation, collaboration, and what he uniquely calls, student-led training.

BOOKS AUTHORED BY RUFRAN FRAGO

  1. Risk-based Management in the World of Threats and Opportunities: A Project Controls Perspective.ISBN 978-0-9947608-0-7.Canada
  2. Plan to Schedule, Schedule to Plan.ISBN 978-0-9947608-2-1.Canada
  3. How to Create a Good Quality P50 Risk-based Baseline Schedule.ISBN 978-0-9947608-1-4.Canada
  4. Schedule Quantitative Risk Analysis (Traditional Method).ISBN 978-0-9947608-3-8.Canada
  5. RISK, What are you? The Risk Management Poem: Children's Book for all Professionals.ISBN 978-0-9947608-4-5 (Canada)
1 comment
36 views

Permalink

Comments

05-29-2021 10:02 PM

Though theoretically ML should correspond to P50. Most Likely (ML) duration may or may not always be at P50. P value will be be dependent on organization / project's risk tolerance, e.g. some may use P80 for ML , others may go with M35.