If you try to do an experiment or a calculation and it doesn’t work out, should you tell other researchers about it? Or just move on to something more promising as quickly as you can? Philip Ball ...