Sometimes the problem is presented in such a way that it seems that it is the shear existence of evil, any evil no matter how small, that creates the problem for theism. And, I suppose that naively we may think that a universe created by an all-powerful, all-loving, and perfect being would have to itself be perfect and thus one in which evil is non-existent. If we are at all swayed by this intuition, then it might appear that the task for the theist is to explain why God would allow any amount of evil to infect His creation.
I think that the whatever argumentative force the free will defense has results from this way of understanding the problem of evil. To most minds, it is at least as obvious that free will is of tremendous value (and that thus an omni-benevolent God would grant it to His creation) as it is that God is opposed to evil. Granting this, it is a short step to the realization that free will necessarily entails that (at least some) people will (at least sometimes) choose to engage in activities that have bad consequences. Thus, God, if He is indeed all-loving, cannot help but create a world in which evil is at least a possibility (since if He is to create conscious beings at all, He must, out of concern for their value, grant them the option of choosing to do wrong).
So long as we understand the problem of evil narrowly, as the need to explain why God would create a less than perfect world, the free-will defense gives us a plausible resolution of the problem. However, a different and stronger version of the problem points not to the existence of evil in general but to specific evils (van Inwagen would call this the distinction between global arguments and local arguments). When we consider certain cases of horrible evil (especially cases involving children) it is far more difficult to explain why God would allow them to occur. Sure children are often the victims of (bad) free choices but, given the arguments of my two previous posts, God can prevent much of the negative consequences of an act without interfering with freedom.
Now if we assumed that God needs to prevent all evil, then the level of intervention required would, as Dilley suggested, undercut the value of freedom. But if instead we believe that the real problem of evil is to explain why God would not prevent horrific cases of evil (such as the holocaust, the slaughter of innocent children, etc.), then, since preventing these instances would not involve constant intervention (and thus not undermine human freedom) the free will defense seems entirely impotent as a solution. God could have prevented the suffering and deaths of many of the victims of history’s great villains without infringing upon freedom.
I think that the real strength of the problem of evil lies in its appeal to such horrible instances of evil (think of examples like Ivan Karamazov’s description of the child’s encounter with the Turks). Against this understanding, the free will defense is ultimatey a non-starter as a solution to the problem of evil.