[NEAS: The candidate uses discrete changes in both regression coefficients: the decay rate and the inflation rate. With no dummy variables, she observes a V-shaped pattern in one residual plot and an upside down V-shaped pattern in the other residual plot.]
This candidate uses the project template on regression analysis for loss reserving. He shows the regression results and the residual plots for each model with low (or zero) vs high stochasticity. If you have trouble with interpreting residual plots, use the procedure in this student project. First do the analysis with σ = 0 to see what to expect. Then re-do the analysis with σ = 0.10, to see how stochasticity affects the plots. Then do the analysis with σ = 0.50 for the student project.