I think evolutionary theory can make good general predictions. For example: If two groups of the same species find themselves in complete geographical isolation from each other, they will inevitably diverge (in genetic character). We know that drift alone would ensure that.
So, we should find a difference between British and Irish hares, known to have been separated for 8/9,000 years since sea levels rose after the last ice age, and sure enough, there is a family difference.
Often, a creationist angle seems to imply that because the specific course of future evolution cannot be predicted, then the theory is not predictive. But no science can predict the future of the planet. How can we know about that black hole that we're moving towards that'll start sucking the solar system into itself in the year 2148? Or about the giant asteriod that'll hit us the year before? We can't even predict the next ice age.
So, I think that the "non-predictive" criticism of ToE is bollocks, basically, and due to a misunderstanding of "predictive" which expects crystal ball type magic.