Bear in mind that the relationship is only approximate, since they are measuring different things. (Sorry if that's obvious)
e.g. I believe the numbers at the low end can change differently if you switch to a low carb diet, generally the hba1c goes down, but some people find the low carb diet raises their fasting bg. Also as Brunneria said, fasting glucose levels can vary for lots of reasons.
This guy said his fasting glucose went down a few days after he stopped eating breakfast, but I've seen other people say the opposite.
I used to just assume that hba1c was by far the superior measure, but more recently I've read that it can vary with some blood parameters* so some people consistently have higher/lower hba1c than an "ideal" test. Maybe they'll produce an adjusted version of the number one day.
Lower is better, for all the tests, but how low is required for best or reasonable health is uncertain.
* Sorry, I don't remember what parameters, iron levels or red blood count or cell size or something ...