While we have growing numbers of massive star observations, it remains unclear how efficient the key physical processes such as mass loss, convection and rotation are, or indeed how they impact each other. We reconcile this with detailed stellar evolution models, yet these models have their own drawbacks with necessary assumptions for 3-dimensional processes like rotation which need to be adapted into 1-dimensional models. The implementation of empirical mass-loss prescriptions in stellar evolution codes can lead to the extrapolation of base rates to unconstrained evolutionary stages leading to a range of uncertain fates. In short, there remain many free parameters and physical processes which need to be calibrated in order to align our theory better with upcoming observations. We have tested various processes such as mass loss and internal mixing, including rotational mixing and convective overshooting, against multiple observational constraints such as using eclipsing binaries, the Humphreys-Davidson limit, and the final masses of Wolf-Rayet stars, across a range of metallicities. In fact, we developed a method of disentangling the effects of mixing and mass loss in the ‘Mass-Luminosity Plane’ allowing direct calibration of these processes. In all cases, it is important to note that a combined appreciation for both stellar winds and internal mixing are important to reproduce observations.