Suppose that in heaven there will everyday be a choice that we can make, and that we can make it without its coming to be a habit that we choose in a particular way, so that there is some approximate independence between the choices. For definiteness, let's suppose it's a choice of what to have for breakfast. For any infinite sequence B of possible breakfasts (pancakes, nothing, waffles, waffles, egg sandwich, ...) let EB be the proposition that B is the sequence of breakfasts I will eat. Also suppose that I am in heaven. Then the probability of EB is zero or infinitesimal, for pretty much the same reason that the probability of any infinite sequence of coin tosses has zero probability.
Given classical theism, God can just tell which infinite sequence B of possible breakfasts will be had by me. But given open theism, all he has are these probabilities. So for every infinite sequence B of possible breakfasts, God assigns zero or infinitesimal probability to EB, and hence assigns a probability of one or one minus an infinitesimal to ~EB. Say that an agent is nearly certain of p provided that the agent assigns a probability of one or one minus an infinitesimal to p. Then God is nearly certain of ~EB for every B.
So far it seems to me that open theists are likely to agree with me. But now let's explore some varieties of open theism in connection with this.
1. Suppose, as some open theists (van Inwagen and Swinburne) believe, that there can be a fact of the matter about future contingents. Then one of the EB propositions is true, even though God is nearly certain that it's false. This by itself seems incompatible with perfection. To be nearly certain of a falsehood is bad for you. But God is perfectly blessed. He doesn't suffer from bad states.
2. Suppose there are no facts about future contingents. Then God is nearly certain of ~EB even though he knows that there is no fact about ~EB. It is irrational to be nearly certain of that which one knows there to be no fact about.
3. Interestingly, the above line of thought so far fits well with Alan Rhoda's variety of open theism. Rhoda thinks all claims about contingent futures with a wide-scope "will" operator are false, and EB can be seen as a conjunction of such claims. Thus, God is not only nearly certain of ~EB, but he is actually certain. But now let NB be the claim that tomorrow my breakfast will be other than the first item in B, the day after tomorrow it will be other than the second item in B, and so on. On Rhoda's view, NB is false. But the probability of NB is one or one minus an infinitesimal. So on Rhoda's view, God ends up being nearly certain of something that he also knows to be false.