Even if this is a crappy poll, it proves that this race is very competitive
That's a contradiction.
"Even if the experiment is flawed, it proves my hypothesis" is not a sentence a scientist should ever say, and political science should follow the same rules.
I was referring to the statements above. If you take a presidential election poll and measure its bias, subtract the bias, and apply it to here, you still have Snyder being very competitive. Of course that same bias might not exist, but it gives us a clue. I'm not trying to be a political scientist or anything.