This competition succeeds upon a line of competitions for writer and style\nanalysis of historical document images. In particular, we investigate the\nperformance of large-scale retrieval of historical document fragments in terms\nof style and writer identification. The analysis of historic fragments is a\ndifficult challenge commonly solved by trained humanists. In comparison to\nprevious competitions, we make the results more meaningful by addressing the\nissue of sample granularity and moving from writer to page fragment retrieval.\nThe two approaches, style and author identification, provide information on\nwhat kind of information each method makes better use of and indirectly\ncontribute to the interpretability of the participating method. Therefore, we\ncreated a large dataset consisting of more than 120 000 fragments. Although the\nmost teams submitted methods based on convolutional neural networks, the\nwinning entry achieves an mAP below 40%.\n