Project Risk Management - Advantages And Pitfalls (Page 3 of 6) in pdf

Now when should a risk analysis be performed? Certainly it is not always appropriate

but if there are multiple numerical uncertainties and these uncertainties cause concern, the use

of risk analysis techniques is advisable.

The question often arises as to how a highly sophisticated technique like Monte Carlo

simulation can possibly apply to projects for which much information is not precisely defined

in advance. After all, Monte Carlo simulation was developed for very precise applications

(like the atomic bomb) and, as such, generally requires great accuracy and precision in data

inputs and well-defined probability density functions for the various variables. By

comparison, cost and schedule information on engineering projects is rarely precise.

Estimates are opinions of probable cost, not highly accurate answers.

The reason Monte Carlo approaches can be applied to engineering projects is that

decision-makers don’t expect precise answers. They generally are willing to accept variances

of 5 to 10% when making decisions related to quantitative values involving risk. By

comparison, in highly scientific fields, errors of this magnitude cannot be tolerated. Such is

not the case in engineering and construction. It's a different world. Because it is a different

world and because we are more tolerant of some error in engineering or construction projects,

we can de-escalate the requirements for the user of Monte Carlo.

The risk analyst must understand that the primary concern in the minds of managers is

how large the economic exposure is if the project is authorized. Few managers will accept a

project with a low probability of success and a high exposure. Managers rarely are risk takers.

They are risk averse.

Often the worst case scenario approach is used to estimate exposure. In this approach,

all major variables are estimated at their most extreme unfavorable values and the exposure is

calculated. The results are inevitably horrible. They are the theoretical worst case. They are so

mathematically remote that at best they are useless; at worst they are misleading. Even in a

highly risky project, not every variable will go to its unfavorable extreme value. Some

variables will show little variance from the plan and others may deviate in the favorable

direction. Therefore the actual exposure generally is much less than the worst case scenario.

Monte Carlo Simulation can identify what the actual exposure really is and what the

probability of success is.

If, for example, you have performed a risk analysis and told management that they

have a 80% chance of success but that they have a 45% exposure, what does that mean? That

means that the bottom line decision variable can erode by as much as 45% of the target value.

"Why?" will be the next question they'll ask. "Why will it be that way?" They will

want to look at a ranked list of risks and opportunities so that they can search for controllables

and can challenge the management team to come up with alternative strategies and tactics.

These alternatives can then be tested in the Monte Carlo environment to arrive at an optimum

solution.

Monte Carlo is not only a great tool for evaluating a current plan; it's a marvellous tool

for evolving a better one.

Contingency

In any estimate or project plan the estimators always include an item at the end of the

estimate for contingencies. A contingency allowance is necessary because uncertainty exists

in the estimating data and assumptions. The costs cannot be defined precisely when the

estimates are made. To account for this many companies tend to use an allowance of about

10% for contingencies in their estimates. In some cases, the 10% allowance is company

policy. All their estimates include a 10% contingency. This is totally fallacious. The amount

Project Risk Management - Advantages And Pitfalls Page 3

Related Articles

Related forms

Related Categories