Using the sum of the Expected Monetary Values of key risks is one way to do this. Even in the absence of historical data, a team can map qualitative probability assessments to percentages using a rule of thumb (e.g. low = 10%) and estimate the cost or schedule impacts.
For example, if I have two key risks, one of which has a low probability of occurrence and an impact of $10K if it occurs, and a second which has a moderate probability of occurrence and an impact of $20K if it occurs, and I used 10% for low probability and 25% for moderate probability, then I could include the aggregated EMV of $6K as part of my reserve analysis.
You need statistical methods to calculate probability which is one of the component needed to calculate contingency reserve but not to calculate contingency reserve by itself. Except you use your organization knowledge management environment to find data from the past and make some predictive analysis. Saving Changes...
I would add a three point estimate to the risk impacts, multiply them by the probabilities and simulate the schedule.
For instance, if you have a risk with a probability of 10% and a impact of $ 10,000 and a risk with a probability of 30% and a impact of $ 5,000, you can transform the impact in a triangular distribution, using a standard measure of uncertainty (as AACEI's maturity scale, for example). Then you can simulate in an Excel based software and have an estimate for the project cost or deviation from the baseline. Once you have that, I would use one of three things:
- The medium value (average);
- P50, if it is bigger than the average;
- P80 or P90, or whatever is the level of safety you want to have in your project.