You are you after all. But you are also still just a human being, subject to their quirks, foibles and limitations too.
> We just don't have enough knowledge to fast forward the simulation?
Well we could simulate a trillion possible outcomes because we're talking a simulation that involves a trillion variables. Which one would definitely occur we could only be sure in hindsight.
Maybe just like how computers can't truly generate a random number, we can make it appear virtually random. Maybe freewill is somewhat similar in that regard, that there's quite a bit of freedom in a lot of choices, but overall it's not absolutely free will that we're operating on.
the Free Will (TM) state machine uses this constantly updated training data to bias its decisions throughout its shelf life. so there are rules (like you stated) but there is also a huge set of data that alters the rule's parameters, to such extent that some times, new rules are created and old ones get discarded.
free will in itself will never be provable.