The Simple Reason Nate Silver's Final Four Prediction Failed Just Like Everybody Else

With the launch of Nate Silver’s new ESPN-backed just prior to the NCAA Tournament, the site’s first big test was its NCAA Tournament Predictions model.

Before the NCAA Tournament, FiveThirtyEight’s model named the favourites to reach the Final Four, including two 4-seeds, and a lot of people used the model when filling out their own brackets.

But like almost everybody else, not only did the model fail to predict the Final Four, it didn’t even come close.

Of FiveThirtyEight’s Final Four teams, Florida, Louisville, Michigan State, and Arizona, only the Gators were still alive after the first four rounds. However, nobody should be surprised by this.

FiveThirtyEight’s numbers actually showed that its Final Four prediction was almost certainly not going to happen.

According to the FiveThirtyEight model, Arizona had the best chance of reaching the Final Four with a 42.4% chance of winning the West regional. However, that also meant that there was a 57.6% chance they would lose one of their first four games.

Florida had a 41.1% of reaching the Final Four, followed by Louisville (38.1%). Michigan State (24.0%) was the longest shot and actually had a 76.0% chance of not making it to the Final Four.

While three of the teams were given a better than 1-in-3 chance of reaching the Final Four, basic probability shows just how little chance there was for all four teams to make it all the way.

Based on FiveThirtyEight’s model, this Final Four combination had a 1.6% chance of happening.

Even if we eliminate Michigan State, who was in the wide-open East regional, the chances of the other three teams making it to the Final Four was still only 6.6%, or slightly better than a 1-in-20 shot. That is still very unlikely.

Modelling is much easier when there are a limited number of possible outcomes and those outcomes can be simulated beforehand (e.g. political polling).

But while the predictive model in something as complex as the NCAA Tournament can tell us the outcomes that are more likely than others, it is impossible to narrow down those possibilities to something that is likely to happen.

Just ask Warren Buffett.

NOW WATCH: Briefing videos

Business Insider Emails & Alerts

Site highlights each day to your inbox.

Follow Business Insider Australia on Facebook, Twitter, LinkedIn, and Instagram.