But his main thesis, grossly oversimplified, is that when predictions tend to be wrong, it is because those making them do not distinguish between the signal and the noise -- between information, data, and events that are necessary building blocks toward accurate prognostication and other data, information, events and commentary that may appear relevant to some but essentially lead one astray. This is a good framework for looking at the early predictions for the 2014 election.
The current consensus in Washington is that 2014 will be a Republican election -- that they will gain some seats in the U.S. House, that they have a realistic chance of recapturing a bare majority in the U.S. Senate and that they will continue to enjoy a sizable edge among the nation's governors and hold their own in the state legislatures.
This common wisdom is based on a series of premises, many of which are more noise than signal.
Among them: A historical record of a two-term president's party losing legislative and executive seats in almost all previous mid-term elections following his re-election; polls showing a majority opposes the president's signature achievement, the Affordable Care Act (ACA); other polls showing Obama's favorability rating at low ebb; missteps in the roll out of the ACA and promises made that citizens would be able to retain their existing health care policies if they liked them that were inaccurate and could not be kept.
snip
We do indeed need to vote, frazzled.