I think that determining a spread that I can work well with goes a long way toward making or breaking my "accuracy" rate. (And even when I find a spread that seems to work well for me intuitively, as the others have said--no one is always right, anyway!)
I would have trouble using the spread you linked to for a couple of reasons. First, I have a hard time determining what single card answers a Yes/No question (the last position of that spread.) For those who have a fool-proof method of answering Yes/No questions with one card, that wouldn't be a problem...but I am not among those readers.
I also should think that the seeker would already know how s/he feels about the other person (Position #2), as well as how s/he sees him or her (Position #4)--which are two out of the five questions in the spread. So, I don't understand what benefit the reader gets from asking them. Therefore, again, that spread would hinder
me.
I'm not trying to pick on the designer of that spread...if it works well for her or for you, then that's great. I guess my point is that some things as seemingly simple as choosing the right spread for the question can challenge your "accuracy rate", too.