SIDEBAR: INTERACTIVE MONTE CARLO
Monte Carlo simulation is a deceptively simple but powerful tool for risk analysis.To try the tool: go to
SIDEBAR: The Shape of Risk
Even without any numbers, the basic probability curve can convey plenty of information about the risk it describes. Here's a cheat sheet for deciphering probability curves
The Basic Curve The basic probability curve looks like an anthill. Here, the X axis represents potential outcomes from worst to best, going left to right. The Y axis represents the probability of those outcomes, from lowest to highest, going bottom to top. The highest point on the curve indicates the most likely outcome of the risk. The best case falls at the far right and worst case far left, both with the lowest probabilities of occurrence.
A steeper, narrower curve (the red line) represents more certainty about the outcome, since more potential outcomes fall in a smaller range. A low, broad curve (the blue line) represents less certainty about a risk's potential impact on a project. With this understanding, you can determine the likelihood of potential risk outcomes with a quick look at a distribution chart.
The Optimistic Curve While steepness of the curve indicates certainty, its tilt describes relative outlook. A risk distribution that tilts to the right represents a more optimistic outlook, since the higher probability results are closer to the best possible outcome.
The Pessimistic Curve On the other hand, a curve that leans to the left shows a more pessimistic view of the risk, since there's more probability that the outcome will fall on the worst-case side of the spectrum.
SIDEBAR: The Five Universal Risks to Software Projects
1 Schedule flaws Either an error in the original schedule or an error in the way the project is run can affect its timing.
2 Requirements inflation This happens when what is needed from the project changes during development.
3 Staff turnover When key people leave during a project, it can have a serious impact on continuity and schedule.
4 Specification breakdown Anything less than complete agreement on project specifications can be fatal.
5 Underperformance Substandard work by anyone on the development team will affect project quality.
SOURCE: WALTZING WITH BEARS
SIDEBAR: When to Use Which Tool
Both Monte Carlo and decision tree analyses are powerful tools, but each has its particular strengths. Monte Carlo simulations are good for accounting for multiple risks occurring simultaneously. Decision trees excel at analysing sequential risks compounding over time. Given those frameworks, here's a look at several scenarios and whether you would be better off rolling the dice or climbing the tree.
Decision Tree
- Decision based on monetary value
- Sequential decisions required
- Few variables or low probability variables that are easily calculated
- Analysing two possible decisions against each other
Monte Carlo
- Decision based on criteria other than value, such as a schedule
- Decisions involve one variable
- More than five variables in complex environment
- Analysing an entire portfolio strategy
SOURCE: RISK AND DECISION ANALYSIS IN PROJECTS BY JOHN R SCHUYLER
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.