We will present progress on our efforts to validate the avalanche model in JOREK by modelling a JET experimental case (pulse #95135) in which an RE beam was generated by a massive injection of Ar gas. The JOREK simulations are 2D (axisymmetric) and use the RE fluid model developed by V. Bandaru. The objective is to show, if possible, that one can obtain a good match with experimental data only when using the correct avalanche gain and not when using a falsified gain. In other words, we would like to show that we are doing real validation and not a mere 'fancy fit'. We have started using Bayesian optimization tools and the synthetic synchrotron diagnostic in JOREK (not yet combined together) to help us in this task.