Decision Tree

A diagram, chart or model which helps support the of decision making, providing a sequence of actions and outcomes, including the likelihood of each option and value of the resultant outcome.

Decision trees will typically follow an IF and THEN pattern: IF this is selected THEN this will most likely occur. It can, however, also incorporate other features such as conditional probabilities, whereby the probability of an outcome occurring is dependent or conditional on another event having occurred or being present.

For example, the likelihood of a football match being cancelled due to waterlogging, is conditional on, for example, significant features such as the type of pitch construction, the current soil condition (e.g. degree of compaction), recent rainfall figures an forecast figures for the day of play as well as preceding day – along with other features.

IF the pitch is a sand based THEN the likelihood for waterlogging (in very general terms, but this would be refined by including a range of other relevant variables) in a region might be 2%, whilst for a soil-based pitch it might be 20%.

However, if conditional probability is incorporated into the decision tree, and the degree of compaction is high, then for the sand pitch this might rise to 5%, yet for the soil-based pitch this might be 40%.

The more options which are included within a decision tree might improve overall decision making, but it may also result in a more confused picture because many features may not interact as anticipated, creating an illusion of accuracy in outcomes.

With the many variables and interactions between variables in a turfculture situation, the relationships, and in particular the strength of the relationship, between the many variables would need careful consideration in arriving at an informed probability value for any selected option. In practice this will most likely be able to identify a likely trend in the direction of travel for the selected options rather than provided hard and fast data.