Doesn't matter IMO. What matters is the change-in-percent. If Florida's "undercount" is 300% higher this week than last week, that's all the evidence we need, as long as the "undercount" remained consistent.
The exact number isn't important. Its the change in numbers we can look at and rely upon. If you don't trust the numbers, you can still trust the change-in-numbers.
Its not important to get exact numbers. Its more important to remain consistent in your reporting through these crisis. We care about apples-to-apples comparisons within the state, not against different states. A 300% rise is a 300% rise.