The tradition of gravity models is in the analysis of trade flows with market size and geographic or economic distance as core variables. Both these variables can be important determinants of FDI, too. However, when such models are used to explain FDI, there can be differences in the mode of operation of these variables so that the interpretation can become uncertain. Market size can reach beyond the host country and distance can be an incentive as well as an impediment to FDI. In the present paper, we use gravity-type models in order to assess the level of German FDI in CEE countries, distinguishing between the four nearby core countries and the other six EU accession countries. Estimates are done both on the basis of an in-sample as well as an out-of-sample approach on the basis of FDI in 81 important host countries. From the various specifications it becomes obvious that (1) beside the size of the host country, the market potential of neighboring countries is an important determinant of the regional structure of German FDI, and (2) the costs of operating at a distance seem to be weighted higher than the advantage of being close to distant markets. Geographic distance seems to be more important than various measures of economic distance. On the whole, gravity-type models seem to be appropriate to explain the regional structure of FDI. However, the results should be interpreted with caution given the extremely wide confidence intervals of the estimates which is a feature of former studies, too.