The erro surely comes from (a) treating probilities as some kind of absolute cause, when they are merely a measure of ignorance and, (b) relatedly, abstracting real situations as numerical items.
In the case of Cornelius Hunter, knowing exactly the entire causal chain would give a probability of his birth as 1.
On the other hand, knowing the entire causal nexus of the growth and existence of a tree leaves the probability of its reciting Shakespeare as 0. Why? Because in the first case there is a valid causal chain and in the second there is not.
Therefore, some biological question such as “the probability of a self-replicating molecule forming” means nothing, whether or not it is known to have happened, except as a measure of our ignorance of what causal chains are available, or have been followed in the past.
From another angle, to treat any rare event as a question of probability, in a vacuum, is to make a metaphysical assertion that “chance” is some kind of cause.
In Joshua’s 3rd case, there is really no probability involved at all, once one believes that such providence exists. The events that occur have an “absolute” (from the viewpoint of God’s knowledge) probability of 1, and those that do not, of 0.
Did God determine that Corneliues Hunter should exist? Yes - probability, 1.
Did God determine that this tree recite Shakespeare? No - probability 0.
As soon as you abstract from “Cornelius hunter” or “this tree” to probabilities, you have lost sight of concrete realities, and are dealing with artificial generalities (Cornelius is not “an example of random fertilisation,” and “this tree” is not “a tree.”