>In 2000, both Gore and Bush were always the frontrunners, and then they both won the primaries in landslides. I believe they both won every state, except Bush lost New Hampshire to McCain.

I think the question is: does the ability of pundits to predict elections correlate or anticorrelate with voter participation rates?

Or in simpler terms: do pundits predict well when primaries are democratic and reflect the will of the people, or when they're isolated, low-information games that mostly reflect punditry itself?