Probability of exactly forecasting the electoral vote in the last three elections

11 Feb

Richard Charnin
Feb. 11, 2017

I was asked to calculate the probability of my exact forecast of the Electoral Vote in the last three elections (365,332,306). It was a combination of experience and luck. I do not expect to exactly forecast the EV in 2020.

Note that the following calculation is just an approximation.

Assume the following:
1) the probability of Obama winning in 2008 was 0.95; it was also 0.95 in 2012. The probability of Trump winning in 2016 was 0.05.
Therefore the probability of forecasting all three winners correctly is
P1 = 0.045 =.95*.95*.05

2) the winning EV is in the 270-370 range.
The probability of exactly forecasting the EV in a given election is 0.01. The probability of exactly forecasting the EV in all 3 elections is 1 in a million:
P2 =.000001 = 0.01*0.01*0.01

Therefore, the probability of forecasting the winner and the EV in the three elections is
P3 = P1*P2 = .045* 0.000001 or 1 in 22 million.

To put it another way, forecasting the electoral vote exactly in three successive elections would be expected to occur just once in 22 million elections (88 million years).

Posted by on February 11, 2017 in 2016 election

2 responses to “Probability of exactly forecasting the electoral vote in the last three elections”

1. February 14, 2017 at 9:55 pm

Mr Charnin. I am a high school math teacher, and one of my classes is Statistics. I am fascinated by your work and wanted to know if you had any suggestions on how to use some of this data for confidence intervals or hypothesis testing. Raw data, and maybe where to find it or which subjects to concentrate on. I read your book as well and would love to be able to use this data with my high school seniors.
Thank you very much for your time and I appreciate so much what you are doing. How can we better get your message and facts out to the greater public? Thank you for anything you might have time to share

• February 16, 2017 at 7:33 pm

David,

Thanks for the comment.

I use the 95% confidence interval which is a function of the standard deviation and the margin of error.
You have one of my book(s) (which one?), so you have all the stat related info to use in your class.
It would be great if your students each had a copy (Kindle or paperback).
Or you can direct them to my blog. https://richardcharnin.wordpress.com/