i guess it depends a bit if you want to do "any old PhD", that is, if you really want a PhD and the topic is only secondary, or if you are interested in that specific project. In other words, is it the interesting project that makes you want to do a PhD, or is it wanting to do a PhD that led you to that project. In the former case, maybe a backup PhD is not such a good idea, or, only if you find another one that you really find equally interesting. In the latter case, you should definitely go for a backup since otherwise you might end up without a PhD.