The principle of alternate possibilities was debunked by the Frankfurt cases back in the 60s. An unrestricted ability to do otherwise is simply not relevant to moral responsibility.
I've elaborated on why the control I describe is sufficient below:
1 . Compatibilists don't assert that determinism is true, they assert that determinism doesn't matter. Which means if the world is non-deterministic, Compatibilism still applies. This is not true of hard determinism.
2. Hard determinists have no way to assign blame in a way that disallows morally repugnant conclusions. If someone punches you in the face unprovoked, how does a hard determinist argue that the unprovoked attacker is the person requiring rehabilitation? Why shouldn't the law be changed to allow unprovoked face punches? Why shouldn't you be rehabilitated to learn to duck faster? There is no moral blame, so anything is permissible.
To #2 the reason to continue mostly as before is because it is what we intuitively feel is fair and that minimises harm. The idea you need to blame someone to do this is a typic religious argument against atheists - we do not need blame to have a feeling for what we consider right and wrong. A belief in determinism shifts those feelings.
It's very simple to reconcile - the only change is that punishment for the sake of punishment rather than to minimise harm is not justified if people have no agency.
To a lot of people that is a belief they come to anyway without a belief in determinism, because of the number of moral systems that includes forgiveness as a central concept.
Consider the prison system in Norway vs US for example. Norways prison system was awful up to the 60s where a reform movement won support for a focus on rehabilitation over punishment.
For you to have "control" you would need to be able to produce different decisions from the same starting point, or there is no control - just determinism.
And those differences needs to not just be purely random, or there again is no control.
There then needs to be some mechanism that makes that choice. If you can not derive that from cause and effect of interactions of the physical aspects of the brain, then that is dualism. But that does not resolve the free-will issue:
That just means that there is an unknown decision mechanism, that needs to be subjected to the same question of how it makes the decision in a way that is not deterministic, and not random, yet produces outcomes that differ when re-run with the same state.
Free will basically would require some third distinct category between determinism and randomness, that infuses some sort of agency into a process (and that is not just a combination of the two, but qualitatively different).
I've yet to see someone come up with a description of this "something" that is not just entirely obfuscation.