In this paper, I want to investigate the variations in the translations produced by a group of students translating the same source text. The aim of the study is to investigate the degree to which it is possible to empirically find traces of Third Space characteristics of legal translation of court decisions between Danish and German. A corpus of 22 translations is investigated for patterns followed by the students, with a special view to assessing the degree to which they choose similar renderings of the source text elements. A central result of the empirical study is that there is a considerable degree of variation across the studied translation students and that a general distinction between strategies guarding source text complexity in content and form, on the one hand, and strategies making the rendering of the source text in the target text culture more accessible for target text readers.