Hypothetical horse race polls all face the same problem: differential turnout. Asking the same people, “if the race were between x and y? What about x and z?” Makes the assumption that the same people who would vote in the first race would also vote in the second.
Highly doubtful. El-Sayed and Stevens poll within several points of one another against Rogers. It is virtually certain that voters will come home by November. This is wishful thinking.