The Verification Instrument for the Direct Assay of Reactors at Range (VIDARR) is a reactor monitoring tool based on solid plastic scintillator. Reactors produce c. 10^20 anti-neutrinos per gigawatt (thermal). These anti-neutrinos can then be captured by protons in the detector via inverse beta decay (IBD). This produces a distinct double coincidence signal in the VIDARR detector. The GEANT4 detector simulation is used to estimate the improvements made by the ongoing upgrade to VIDARR. The new geometry of the detector sees a significant improvement in containment of the Gadolinium gamma cascade, showing c. 25% more high-containment events. A traditional optimisation of a neutron trigger is conducted using simulated detector results with two different thresholds and is compared to a machine learning approach using a Support Vector Machine (SVM). Which improves neutron signal efficiency by ~ 5% and purity by ~10%. Finally, a data driven dark noise model is added to increase the simulation accuracy and to fold in electronics effects. This is further augmented using a more accurate background model.