At some point in the game yesterday (which I didn't watch), prestigious forecaster Nate Silver gave the Patriots a 0.5% chance of winning.

Offhand I'd say that the "prior probability" that a pundit for a prestigious newspaper is a charlatan is 20%.

Then that pundit tells me that the probability of event X is 99.5%, and he's wrong. What now is the probability that he's a charlatan?

His tweet:

"That Patriots drive took another 5:07 off the clock and actually dropped their win probability from 1.1% to 0.5%:"

Just for fun with numbers, here's a simple exploration of the fourth quarter:

Again, the score at the beginning of the 4th quarter was Falcons 28 - Pats 9. The Falcons had thus "averaged" 9 1/3 points per quarter, but had scored 21 pts in the second quarter, so let's make an assumption they could possibly have scored 14 in the 4th - again, #1 offense in the NFL.

So here are scoring possibilities for the Falcons in the 4th (leaving out safeties as unlikely, and 2-pt conversions as unnecessary since they were in the lead): 0, 3, 6, 7, 9, 10, 12, 13, 14. In other words, they could score 0 pts, or just gotten a field goal, or two field goals, or a touchdown with EP, and so on up to 14. For a total of 9 outcomes.

The Pats have to, at a minimum, tie the Falcons, so including 2-pt conversions, the Pats might score: 0, 3, 6, 7, 8, 9, 10, 11, 12, 14, 15, 16, 17 points, for a total of 13 outcomes. 13 outcomes

Now just create a simple 9 x 13 matrix, and only 1 cell in the matrix has the Pats getting a tie that goes into OT - all other cells are a Falcons win in regulation. So there's a naive model that gives a
0.85% chance of winning for the Pats at the start of the 4th quarter.





Speak your mind

1 Comment so far

  1. Marion D S Dreyfus on February 7, 2017 4:45 am

    Nate silver was entirely wrong on predicting a Trump win, too, so maybe his predicting career might take a sabbatical until something more sussible comes his way.



Resources & Links