Who is likely to win today's Super Bowl? I ran two very simple models to find out. Both were based on the number of points for (PF) and points against (PA) during the regular season plus the playoffs. Both models predict a win by the New England Patriots over the Seattle Seahawks.
The first model is very cautious, giving New England a 51% probability of victory. The second model is less cautious, giving an 81% probability of a win to New England, as well as a predicted margin of victory of 2 points. We shall see later today. Play ball!
Update, the morning after: I ran a third model, again predicting a margin of victory of 2 points. However, this model concludes that the probability of a New England victory is 54%. This model seems the most sound, as it is based on only the two teams' records during the season, as opposed to the second model, which relied on the records of all the teams in the league. It seems to be borne out by the closeness of the game.
Yet another reminder that statistics works best with large sample sizes. When n = 1, you are taking a big chance!
The first model is very cautious, giving New England a 51% probability of victory. The second model is less cautious, giving an 81% probability of a win to New England, as well as a predicted margin of victory of 2 points. We shall see later today. Play ball!
Update, the morning after: I ran a third model, again predicting a margin of victory of 2 points. However, this model concludes that the probability of a New England victory is 54%. This model seems the most sound, as it is based on only the two teams' records during the season, as opposed to the second model, which relied on the records of all the teams in the league. It seems to be borne out by the closeness of the game.
Yet another reminder that statistics works best with large sample sizes. When n = 1, you are taking a big chance!
RSS Feed