Here are two forecasts, both from 538 (as of 6:30am Pacific, 2016-10-19).

First, we see that Hillary is forecasted to win the election. Also, we see the history of this forecast (more on the history of this forecast below…it’s actually the point of this post).

Second, we see the forecast for the World Series. Note – and you’ll have to take my word on this (or comb through Google’s cached web pages) – that the Cubs, yesterday, were forecasted to have a 38% chance of winning the series, tied with Cleveland which also had a 38% chance of winning.

538-polls-only-19-october

 

mlb-predictions-19-october-am

 

So, here’s the question: Are forecasts useful for anything?

Here are three more questions that help to explain the above question:

If we consider a forecast to be static (not changing in time), and the forecast has “error bars” (that is, is, say, 95% accurate) what does that mean? and is it useful?

If a forecast changes in time, is it useful?

(What the word “useful” means is important in answering the questions. The answer to the questions will change dramatically depending on what specific utility is sought. I ask the questions from this perspective: by useful, I mean to say that an individual would make an exceptionally important (life changing perhaps) decision based upon the forecast. For example, a middle-class individual might take out a second mortgage on their home to place a $250,000 bet on the outcome of the World Series.)

-JD Cross

 

Advertisements