CU Electoral ‘Prediction’ Debunked
I am a little late to this party, just wanted to clarify my position on Bickers and Berry’s model suggesting that Romney is likely to win in a landslide versus Obama.
“For the last eight presidential elections, this model has correctly predicted the winner,”
No it has not. The model was just created… how has it predicted anything?
It backed into past results, never once predicting an actual race. Big difference. To make an analogy, it would be similar to finding a trait of every Super Bowl winner and then applying the measure to the upcoming Superbowl saying that the model has predicted every Super Bowl winner.
Just so we are clear, their map has Obama losing every battleground state. If we assumed that each battleground state was a coin toss 50/50 chance of winning for either candidate, which it is not (Obama’s odds today of winning many of the BG states greater than 50/50), but just to give these gentlemen a benefit of the doubt… The odds of this event happening is 1 in 512 or about 20% of 1% with respect to likelihood.
If you would like to experiment at home, start flipping a quarter in the air. Designate Romney as ‘heads’ and see how long it takes to get 9 heads in a row.
“What is striking about our state-level economic indicator forecast is the expectation that Obama will lose almost all of the states currently considered as swing states, including North Carolina, Virginia, New Hampshire, Colorado, Wisconsin, Minnesota, Pennsylvania, Ohio and Florida,” Bicker said.