The problem with CNN is that they were actually right about the numbers and wrong about how to interpret them. Almost all of the mainstream polls were about na

The problem with CNN is that they were actually right about the numbers and wrong about how to interpret them. Almost all of the mainstream polls were about national popular vote, and predicted that Clinton would win it (she did) and within their margin of error (about 2 points lower than they predicted). The problem is that they construed winning the national popular vote as winning the election, and while they're correlated, one doesn't entail the other. So while their prediction about popular vote turned out to be correct, their misestimation was localized mostly in a few key regions, which narrowly flipped to Trump. And thus we find ourselves in the case where CNN pollsters were right and CNN analysts were bad at explaining it (and made lots of untrue inferences).