Right.
It seems to me that when you deal with the legitimate chance issue, one has to make a call at some point as to where that stands. Poll numbers go up and down. They're particularly going to be the case if you have a party that contests only in one province, for example. It's a relevant question, because the Bloc Québécois is one of those parties. The margin of error goes up when you're dealing with a smaller sample.
Have you thought about how you're going to deal with the timing issue? At some point, the decision can itself make a significant impact on whether or not that party is going to be able to win seats in the next election.