For the hotend i have built i was aiming for a very short melting area and as little nozzle/hotend mass as possible that is beeing heated to melting temperature. The design with the shortest melting area (< 10mm) and the shortest transition area (ca. 10mm) turned out to be the most reliable.