Physically based models, resolved using the finite element (FE) method are often used to model changes in dimensions and the associated stress fields of graphite moderator bricks within a reactor. These models require inputs that describe the loading conditions (temperature, fluence and weight loss ‘field variables’), and coded relationships describing the behaviour of graphite under these conditions. The weight loss field variables are calculated using a reactor chemistry/physics code FEAT DIFFUSE. In this work the authors consider an alternative data source of weight loss: that from a longitudinal dataset of density measurements made on small samples trepanned from operating reactors during statutory outages. A nonlinear mixed-effect model is presented for modelling the age and depth-related trends in density. A correction that accounts for irradiation-induced dimensional changes (axial and radial shrinkage) is subsequently applied. The authors compare weight loss forecasts made using FEAT DIFFUSE with those based on an alternative statistical model for a layer four moderator brick for the Hinkley Point B, Reactor 3. The authors compare the two approaches for the weight loss distribution through the brick with a particular focus on the interstitial keyway, and for the average (over the volume of the brick) weight loss.