Many real estate academics and practitioners agree that measuring real estate risks requires both quantitative and qualitative instruments. Qualitative part scoring models are often used in the industry. Theoretically these are well suited as they condense complex information into one measure, for instance regarding location quality. On closer scrutiny, however, the scoring models in our study lack validity and reliability. As we show in our paper this is largely due to deficiencies in the construction of the models which can be reduced when certain re-quirements are met. The paper begins with an extensive literature overview of the theoretical background, the various applications and the different construction methods of scorings. It becomes clear that the topic is under-researched and that there is a great divide between the academic literature and the practical application of real estate risk scorings. In the second part of the paper we present the results of our (unrepresentative) empirical survey. We analyze the scoring models of four large and broadly diversified commercial real estate investors with a total market value of more than 10 billion Euro. One important finding is that the scores do not show any significant correlation with other risk measures and are thus not ideal for capturing the quali-tative aspects of real estate risk. However, we also find evidence that scoring models can produce meaningful results if empirical-statistical methods are used in addition to expert opinions; thorough validation and calibration are further prerequisites for successfully meas-uring real estate risk with scoring models.