Political scientists have studied pundit predictions and found them to be, on the overall, inaccurate. Indeed, the effect gets stronger as the pundit becomes more popular: "the better known the pundit, the less accurate his or her forecasts." But all this suggests that political punditry has something to do with accuracy. It doesn't. It's entertainment. Just like people who like sports want to be able to watch TV shows about sports and people who like women in bikinis want to be able to watch TV shows about women in bikinis, people who like politics want to be able to watch TV shows about politics. The pundits exist to fill that need. Their role is to make those shows entertaining, so the shows have good ratings, so they can sell time for advertisers, so they can make a profit for networks. Indeed, as everyone knows, a very accurate show with very low ratings will be canceled and a very inaccurate show with very high ratings will become the network's flagship. People tend to stay away from the implications of that, but no one denies that it's true.
