Skip to main content Skip to page footer

Simulating Risk in the Courtroom


The Challenge:

  • Forecast Superior Court demand, where factors influencing future directions are very difficult to predict and there was a large group of stakeholders with differing judgements on these factors. 
  • Develop a useful, actionable set of forecasts while keeping stakeholders satisfied -  a high risk for our client as well as Data Analysis Australia. 

The Data Analysis Australia Approach: 

  • Co-ordinate brainstorming sessions with stakeholders and design and implement a survey to collect expert opinions on different factors affecting Superior Court demand, as well as perceived probabilities for each factor in the short and long term.
  • Run over 500 simulations of demand on the Superior Courts, driven by probability distributions developed from the stakeholder input, as well as other trusted data sources.

The Result:

  • Range of forecasts including probabilities of each scenario occurring.
  • Satisfied stakeholders seeing their qualitative input being directly incorporated in the forecast results.

The Challenge  

Forecasts that rely on policies and practices present a high level of risk to our clients –with so many different possibilities for the future, how do you ensure you’re planning based on the “right” one?  Superior Court forecasts are one such case.  There are changes in offence rates which are affected by Policing, changes in jurisdiction and legislation from Governments that direct cases to lower Courts or out of the system altogether, and there are Judicial practices about how cases are handled and where they’re heard.  All of these can make a big impact on future infrastructure needs.  This in turn poses a risk to Data Analysis Australia:

How can we pool together the (sometimes contradictory) judgements of our stakeholders and produce forecasts that satisfy all parties?

The Data Analysis Australia Approach

Recognising the inherent variability in a range of factors affecting future Court demand, our approach was to generate a probability distribution of possible outcomes for these factors.  By isolating the uncertainties in the individual factors, we enabled stakeholders to focus on them one at a time and to make a reasonable assessment of the probabilities to be estimated.  

To facilitate this, Data Analysis Australia co-ordinated a brainstorming session with stakeholders to first identify a range of factors relating to Court demand to target:

  • Population growth – overseas and internal migration;
  • Offence rates – effect of economic or social conditions and police practice and legislation;
  • Litigation – levels of corporate fraud and personal trials;

Jurisdiction – lower level offences that may move to the Magistrates Courts;

  • Lengths of trials – efficiency of Court proceedings and experience of legal representatives;
  • Listing procedures – accuracy of estimates of trial lengths and efficient use of Court space; and

Courtroom availability – changes in operating hours for Courts and the Judiciary.

We then designed and implemented a survey to gather each representative’s expert view on the probability of different assumptions on these factors proving true in the short term and long term.  For example, representatives gave probabilities in both the short and long term as to whether the population growth would follow, be 1% higher or be 1% lower than the latest official projections for the State.  The survey also highlighted key considerations affecting each factor based on the brainstorming session and other information provided by Data Analysis Australia to guide the decision making process. 

To understand the results we then needed to combine the probabilities to generate the range of possible Courts outcomes once all the factors were simultaneously considered.  Crucially, not only would this range give the ‘most likely’, ‘worst case’ and ‘best case’, but would give probabilities of these worst and best case outcomes, along with other, more sensible upper and lower bounds. 

To achieve this, we weighted each representative’s responses equally, and used statistical techniques to develop a probability distribution around each future possibility.

Combining this with analysis of Court data, Data Analysis Australia ran a simulation to produce a large suite of forecasts covering a range of possible futures for the Western Australian Superior Courts.  Five hundred times we progressed the demand on the system year-by-year into the future, with likelihoods of changes in policies and population each year informed by the probability distributions we developed.

From this suite of forecasts we extracted a median forecast, as well as realistic lower and upper bound scenarios aligning with and adjusted to suit eight different possible Court location scenarios within Perth’s CBD.  Importantly, this method allowed the probability of any potential outcome to be considered, not just the single best and worst case outcomes.

The Result

The use of statistical techniques focusing on understanding probabilities and uncertainties resulted in a range of forecast outcomes being generated, along with the probability of these outcomes occurring.

As we produced a variety of scenarios along with their attached probabilities, the risk of not having a forecast to suit a planning objective was mitigated for our client.

Another critical outcome is that our client had a set of forecasts that were credible, believable and acceptable to the stakeholders involved, as they had all “had their say” and could have the satisfaction of seeing their input feed directly into the work.  

Sensible bounds on the median, low and upper scenarios (not assuming the mid point, best or worst outcome on everything simultaneously, which would be quite unlikely) were based on the actual variation in the probabilities stakeholders had attached to each future direction.  The forecasts were presented along with the current courtroom stock to allow readers to easily identify when facilities were expected to be inadequate and plan accordingly.  The room forecasts took into account other risk mitigation, such as weighting towards multi-functional rooms.

The Data Analysis Australia report included recommendations for increased efficiencies in Court practices that could make the lower forecasts achievable.  This looked at a balance of the over-listing risk (not having enough rooms to meet demand on the day) and under-utilisation (rooms being unused because a case was expected to be heard there but it was dropped).

June 2016