The Weekly Reflektion 19/2023

Getting the team together and organizing a social function with a few games is often a good way to create the team spirit that provides the basis for good performance. Often a few glasses of beer provide a catalyst for discussion. Sometimes even the beer-games can provide interesting insights into risk and uncertainty.

Do you continually update the risk in your project as new information becomes available?

A few years ago, pre-COVID, I participated in a seminar in an offsite location and with some social activities organised in the evening. One of the ‘games’ was particularly interesting as the prize was a free glass of beer. Three boxes, each with a door, were placed on a table. Within one of the boxes there was the glass of beer. Each person was asked to choose a box, but not open it. Another box was opened that did not have a glass of beer in it. The person was then asked whether they would like to change their first choice. Some changed and some didn’t. There were about 30 of us and once everyone had played the success rates were checked. 12 people selected a box and didn’t change when given the opportunity. 4 of them were successful. The other 18 selected a box and did change their selection and 13 of them were successful. It should be noted that as the game progressed the people started to twig the probabilities and the last 6 people all changed their selection.  

When considered independently the two selections seem to have probabilities of one in three and one in two respectively. The result for people that did not change their selection seem to confirm the first hypothesis. However, the results for people changing their selection were significantly better than 50%. The key to understanding the statistics is that the outcome of the first choice has a significant influence on whether to change the first choice or not. If you change this increases the probability of success from one on three to two in three. The key factor is the knowledge that the first box opened doesn’t contain a glass of beer. Try this yourself with a simple event tree analysis with box A, box B and box C.

Event tree analysis (ETA) is a forward looking, logical modelling technique for both success and failure that creates a path for assessing the risk associated with a project or operation. The probability for different outcomes to each event is used to determine ultimate outcomes. The name “Event Tree” was first introduced during the WASH-1400 nuclear power plant safety study (circa 1974), where the WASH-1400 team needed an alternate method to fault tree analysis due to the fault trees being too large. An assessment of probability is one way of handling uncertainty which in turn is a key component of the definition of risk. A final estimate of probability can be used as the input to a decision. Typically, the event tree analysis is carried out once and when the decision has been made the event tree analysis is filed. When we use event tree analysis in our risk assessment process, we should ensure that it is revised each time new information is available as this may affect the outcome.

Thank you to everyone that attended our breakfast seminar onWednesday 3rd May and for your comments and contributions. Plenty new material for our future Reflektions.

Reflekt AS