Wednesday, February 24, 2016

Even I won't say it's over yet, but...

So, I put up a post about why I was looking at polls rather than prediction markets the other day.  And now, the prediction markets seem to be overreacting. PredictWise currently puts Trump's chances at 68%.  Before Nevada, they put it at around 53%.  That doesn't make mathematical sense.  They put Trump's chances of winning Nevada at roughly 90%.  So, once that is out of the way, Trump's chances should go up to .53/.9=.59, or 59%.  For the math geeks, that is called "Bayesian updating."

Trump is clearly the favorite, and the ranks of Type A Trump denialists are shrinking, but we always need to leave room for weird shit to happen.  Either 68% is an overestimate, 53% was an underestimate, or a 90% chance of winning Nevada was an underestimate.  I think there was a little of each going on, but people are finally coming around to acknowledging that Trump is winning.

No comments:

Post a Comment