Case Study: Curtin University of Technology achieves more realistic risk modeling and improved VaR accuracy with Palisade (@RISK & StatTools)

Research at Curtin University of Technology Explores Tools for Teaching Probability and Risk

Curtin University of Technology researcher Darren O’Connell investigated shortcomings in how Australian higher‑education institutions teach probability and risk, arguing that common textbook methods—often inexpensive and easy to use—frequently assume normality and omit widely used statistics and methods. To demonstrate this, he applied Palisade products, notably @RISK and StatTools, to model stochastic price processes for two illiquid securities and to evaluate value‑at‑risk (VaR) approaches under more realistic, non‑normal distributions.

Using Palisade’s StatTools for summary statistics and histograms and @RISK for distribution fitting and simulation, O’Connell found logistic distributions fit the historical returns far better than the normal and performed better in VaR back‑testing (including Kupiec/Christoffersen tests). Palisade’s tools produced tighter rejection zones, less clustering of VaR violations and results more closely aligned with theoretical expectations—yielding more realistic risk estimates, potential reductions in regulatory capital tied up by suboptimal models, and lower training and system costs due to Palisade’s Excel integration.

Curtin University of Technology

Darren O’Connell

Researcher