During the development process of brake discs for commercial vehicles, heat cracks are a frequent problem. Since no profound model to forecast the occurrence of cracks has been presented yet, their prediction is hardly ever possible. The standardized heat crack test puts the brake disc under severe thermomechanical load and therefore forces it into cracking. In this paper, results from a series of heat crack tests on the dynamometer are presented, which provide insight into the hidden processes that accelerate or slow down the heat crack propagation in brake discs. This includes an extensive experimental setup using a thermographic camera, a set of capacitive displacement sensors, a pyrometer, and sliding thermocouples as well as a unique eddy-current heat crack detector that was developed at TU Darmstadt. Continuous monitoring of disc deformation, surface temperature, and crack propagation at high sampling rates provides the base for a new, profound causal model. The model describes a chain of effects including the qualitative and quantitative influence of the surface temperature distribution to local crack propagation rates in connection with local deformation and coning of the disc. Therefore, spatiotemporal patterns of thermal and mechanical distortions of the disc are compared to each other and to the crack pattern in time and frequency domain. The observation of local hardening effects as well as microstructural transformations in regions of high crack growth completes the model. Furthermore, relations between crack propagation and brake disc design are evaluated, which reveal the influence of the cooling channel pin configuration to the crack propagation. Finally, possible future applications of the model are shown, regarding its ability to forecast the cracking tendency of a certain brake disc design during finite element analysis as well as its integration into the development process of brake discs.