Risk management - Simple probability estimation
Simple probability estimation
At present we have considered an UNCERTAIN EVENT as having a likelihood of less than 50% chance of occurrence.
In the earlier example [see 2 pathways] it was 30% and we put it into the plan (proactive planning). We could just as easily have made a judgement that the extra 5 pumps would not be needed and not planned for them directly. We could have incorporated a trigger and a separate action plan with the extra pumps being treated as a contingency.
When we think about a risk or uncertain event how do we arrive at a particular percentage?
Firstly, when you assess an event and you decide it has a greater than 50% chance then it is more likely to happen than not.
This event should be planned for, there should be a proactive set of activities in the base plan to allow for it.
This is pretty obvious really as knowing it is highly likely to occur you would be silly not to plan for it.
Even in this circumstance there are 2 possible scenarios.
- It WILL NOT incur any extra cost (not so likely)
- It WILL incur extra cost (common)
In either case it will need to be planned for.
If the chance of the event happening is less than 50% you may not actively plan for it but should be aware of its impact.
Naturally, you will need to have good process techniques to categorise risks in this manner as we have seen previously [see The risk management process - estimate - size the risks].
These responses will vary from general to specific, from proactive to reactive (contingency) or may even be ignored fro particular reasons.
If an event may save money (even if it has been given a less than 50% chance of success) you should try to drive this event if at all possible.
How do we assess the likelihood?
To try to assess the probability of any event happening to greater than +/- 10% may be futile.
In the real world we often have a good feel of particular events occurring.
This is because often a lot of data has been collected to assess this probability.
In a project you will often try to use previous experience and perhaps gathered data.
However, by their very nature, uncertain events, needing change in planning, have not usually been seen before. So, previous experiences may not very valid.
You can still try to assess a situation based upon experiences in a similar circumstance.
As we have discussed previously estimating is subject to a range of problems [see Simple estimation problems].
Typical estimates of probability might be 10%, 20%, 30%, 40%.
or
1 in 10 (10%), 1 in 5 (20%), 1 in 4 (25%), 1 in 3 (33%).
These assume that you have already made a very good approximation by saying it will be less than / or greater than 50%.
So, in total, there is only 4 to 6 percentages that you will need to consider.