I think the most we can say is there is more than one meaning of free will.
That is not much, is it?
One could also defend a concept of free will: a concept that fits to the usual way we assign free will and responsibility to our fellow humans. Even if the borders can be vague, we are perfectly able to see that certain actions are done free and others were coerced or forced. Calling on certain metaphysical ideas to support that notion is useless and in vain. However, hard determinists stick to this idea: that free will is a metaphysical category, and as such is not supported by the fact that we our behaviour is determined.
From a metaphysical point of view both free and coerced actions are determined, they do not differ in this respect. But they differ in the way they are determined. Free actions are more or less immediately caused by our wishes and beliefs, where in coerced actions other factors come into play, like being in an artificially created situation, wherein we do things we normally don’t.
And to answer George’s question (Lois seems to have lost interest): we should ask to what we assign free will. Obviously we don’t with atoms and neurons. It is on the level of persons. But persons are not metaphysical entities, they are conventional entities. Now we call those actions free, which we feel are caused by our desires and beliefs: we identify with those actions. So even when an action is too fast to be conscious of my decision before doing it, if I identify with it, i.e. recognise the action as being in accord with my wishes and beliefs, it is my free action. This makes flexing the hand in Libet’s experiments just as free as other actions. The studied subjects in the end consciously agreed to joint the experiment, and as they were not forced to flex their hands at exact moments, these flexings were free. But of course they were determined too.