What Sean Carroll is saying is that whilst OR is a useful heuristic, it should never be the final arbiter with regards to a (scientific) theory.
The idea of OR is that among competing hypotheses, the one with the fewest assumptions and equal explanatory power should be chosen, which is what "simplicity" (parsimony) means in this case, it's not a dictionary definition of the word as it is related to hypothesis testing. In this case, the "waste" would be the extra assumptions required to support "God did it". So I fail to see how it is not correct, as the dictionary definition of simplistic is a straw man of what it means within the context of hypothesis testing.
What SC is saying is that when push comes to shove, we can throw OR out of the window if needed.
We value explanatory power -> The ability of a hypothesis to explain phenomena
and parsimony -> how many assumptions a hypothesis makes, an assumption can still be complex but only when justifiable and necessary.
Let's say that we have two theories, A and B.
A explains C,D,E,F and B explains C,D,E,F.
A makes assumptions G,H,I,J and B makes assumptions G,H,I,J,K.
Parsimony would then claim that A is more likely (or the better choice). The reasoning behind this would be that A explains the same amount of phenomena as B, but entails less.
Now, let's look at theories L and M.
L explains N,O,P whilst M explains N,O,P,Q
L makes assumptions R,S,T,U whilst M makes assumptions R,S,T,U
According to explanatory power, we should pick the theory which explains more phenomena. Explanatory power would claim that M is more likely.
This quote directly shows how a misunderstanding creates such a situation in which "God did it" is the simplistic explanation
Why would "God did it" be a simpler explanation?
"God did" it entails everything we know + the auxiliary of supernaturalism. An alternative hypothesis such as naturalism only entails a closed system. Especially in the case of theism, theism is a specific instance of supernaturalism (basic set theory). Theism therefore makes more assumptions than specified naturalism, even if we assume that supernaturalism & naturalism are symmetrical from the outset.
Given all of that, a proper reading of parsimony shouldn't make "God did it" simpler.
OR only favours "God did it" if your interpretation of explanatory power is prior (what "God did it" purports to explain instead of what it actually explains). I disagree with this view as my interpretation of explanatory power is posterior (What "God did it" actually explains). Given the data that we have, "God did it" pretty much fails in explanatory terms.
So "God did it" does not have equal explanatory power against a competing hypothesis such as naturalism. Therefore, even if it is "simpler" (which I doubt as it entails supernaturalism), OR could not imply "God did it" over naturalism in light of relevant observations.
Topic: Why is there something rather than nothing?