Entering Ranges in the Aptage RBD MS Project Template

 

As pointed out by Douglas Hubbard[1] and others, when reasoning about uncertainty quantities, such as the duration of future events, one should capture not only the expected value but also the lowest value and highest value one would reasonably expect. For future durations, this would be commonly use best case (low), expected case, and worst case of PERT estimation. For example, you might expect a task to be 15 days, but it might get done in 10, or it might take 20 days. As a rule of thumb, you’d be 90% confident in the low and high estimates.

 

Configuring the MS Project File

To take full advantage of Aptage RBd you need to configure your MS Project file.

  • Insert two of MS Project’s number fields (Number19, Number20) fields. For usability, you can use rename the fields (using the Field Settings tool – right click on the field header) to something more convenient (see Figure 1).
  • Also insert the %Complete, BaselineStart, BaselineFinish fields.
  • Finally, after the initial plan is entered, remember to set the baseline using the Set Baseline tool in the Project menu ribbon. This will update and persist the task start and finish dates in baseline fields.

fig 1Figure 1

 

You can find the .mpp example shown in Figure 1 here: http://bit.ly/2nS9yir

Entering Duration Uncertainties

 

When using Aptage RBD with MS Project, the users can enter these three-point estimates. As shown along with the Task Duration, we have Low Duration and High Duration fields. These are used by Aptage RBd in Monte Carlo simulations.[2]

 

fig 2

Figure 2. Entering duration ranges in Aptage RBd

The expected value is entered into the standard Duration field, the best case in the Low Duration field, and the worst case in High Duration field. Also, as shown in the figure, these fields are optional, but when used, they handle every risk scenario.

 

Note that Aptage RBd ignores any 0 value in the Low or High Duration fields and uses the default natural uncertainty as discussed in the following section.

Under the Covers

 

Note: In the default case, Aptage borrows from the Agile practice of relying on the Fibonacci sequences for story sizes (1,1,2,3,5,8,13,21,…) The formula for the nth Fibonacci number is the sum of the previous two numbers in the sequence (e.g. 5+8 = 13). Many in the agile community use these so as to not make over-precise prediction. The idea is that the size of an 8 point story is not exactly 8 but is better thought as an (best case, likely case, worst case) estimate with the worst case being the next Fibonacci and the best case being the previous one.  Aptage RBD treats these 3-point estimates as parameters for a triangular probability distribution to be used in a Monte Carlo simulation.

fig 3

Figure 3.

 

This 3-case reasoning is extended from Agile stories to reasoning about durations of tasks. This entails generalizing the reasoning to non-Fibonacci numbers. This is done by noting that the ratio between succeeding Fibonaccis   is approximately 1.62 and conversely the ratio of a Fibonacci with its predecessor is 0.62.[3]

 

 

Here are some examples:

 

Fib(n) F(n+1)/F(n) F(n)/F(n+1)
1 1 1
2 2 0.5
3 1.500 0.667
5 1.667 0.600
8 1.600 0.625
13 1.625 0.615
21 1.615 0.619
34 1.619 0.618
55 1.618 0.618
89 1.618 0.618

Table 1. Ratios of Fibonacci’s

Aptage adopts these ratios as default. So, for example, if as in Figure 1, you enter a duration of 10 and leave the low blank, Aptage RBd will use   6.2 = 10*.62 as a low. If you leave the high blank, Aptage RBd will use 16.2 = 10*1.62 as a high. These are reasonable values when you have little information. If you know better, again as shown in Figure 1, you can override the defaults.

[1] Douglas Hubbard, How to Measure Anything, 3rd Ed., Wiley 2014

[2] Treating of these sorts of estimates as triangular distributions in project management is call PERT estimation.

[3] In fact, the ratio approaches the Golden Ratio of antiquity. See https://en.wikipedia.org/wiki/Golden_ratio

About the author
Leave Comment

Your email address will not be published. Required fields are marked *

clear formSubmit