Thursday, November 3, 2016

My Election Prediction

I am going to make my prediction for the upcoming election:  a tie in the Electoral College, throwing the election into the House of Representatives.  How do I know this?  It's based on a very simple fact.

A successful prediction of a remarkable event is remembered forever, while a failed prediction is soon forgotten.  I suppose one could base an entire career in theoretical physics on this fact...

There was a fun episode of the old TV series, "Alfred Hitchcock Presents," in which the protagonist starts receiving (through the mail, if I remember correctly) a series of predictions of upcoming sporting events.  Every prediction comes true.  Finally, he is so convinced that he bets a large sum of money based on a final prediction.  He loses, of course.

It turns out that the scam artist who was the source of these predictions used a very simple method: he began with a large target audience, and for the first sporting event, he mailed one prediction to half his audience, and the opposite prediction to the other half.  After the sporting event was resolved, he dropped the group of people who had received the incorrect prediction from his mailing list, and continued this procedure with the other half.  So his audience was halved each time, until he ended up with a group of people who were absolutely convinced of his predictive powers. I've often thought that there should be some other (legal!) application of this procedure, but I've never come up with anything worthwhile.  By the way, I am not mailing out 2^538 predictions -- I am making just this one.

No comments: