I think that it's correct to use a "backward-looking" point of view: after all, we have to draw conclusions about the future on what has already happened.

However, not in the conditional sense of "what could have happened", or "could I have made any other choice?". To me, not only is that inherently unknownable, but it just confuses what is a very simple matter, that things happen for previous reasons (causes), and nothing happens without either a known reason (not free will) or unknown randomness (which is not free will either).

So, take any decision process that is claimed to be a "free will" process, and just work backwards. Ask why that decision was made. Either the acting party knows or doesn't. i they know, voila, there we have out determinant reason (cause). A different reason (cause) would have led to a different result.

On the other hand, the acting party might have no idea why the decision was made, so that can hardly be called free will. It just happened as if the decision had fallen out of the sky (randomness).

I really think it's that simple an issue. For any decision process there will be a series of "why" questions to determine how one chose this or that decision. Determinism or randomness (where randomness simply means we don't yet know the deterministic cause due to lack of knowledge).