I agree with the math, only in this case we have experience and reason about our inputs.
For example, I have a production crew that has been previously caught drinking on Fridays. I notice that after several months the quality control reports are trending down on Fridays but not enough to get alarmed about. In this case I have a known variable which will cause a decrease in quality but it is not expected to be present. Because I know about this variable, it should cause concern.
In the instant case, we have an unfounded assumption that the moderator and both candidates will be fair about time.
Fair about time or anything else!