Each week, BRSN sits down for a Q & A session with...
Ivy League Elo III
Ivy League Elo is back with another five years of ratings. The 1963-1967 football seasons have now been charted and are shown above. Also, at the end of the article is the cumulative graph of all ratings that have been calculated. If you want to know more about the calculations or Elo ratings in general have a look at the first Elo article. In this update, no changes were made to earlier calculations or graphs.
During this five year period, the Big Red were a textbook average team. They hovered around a 1500 Elo rating and finished from fourth to sixth in terms of rating each year. The team’s rating peaked at 1567 (their highest rating yet) at the end of 1967 season, indicating the team was on the rise. Other than that peak, not much happened with Cornell in this period.
The teams that ruled from 1963-67 were Dartmouth and Princeton. They collectively claimed four of the five Ivy League and Elo Championships and continued to dominate in ratings. Princeton was able to reach a rating of 1723 in the 1965 season, which is the highest any team has reached so far. These two teams are also the only teams to have reached 1700 in their ratings and were also extremes on the other side of ratings. The University of Pennsylvania set the record for the lowest rating yet with a score of 1267. It was the culmination of an unwinning season that had followed several other poor seasons. However, Penn was not alone; Brown and Columbia both spent their time in the bottom of the rankings.
Let’s take a look at Elo’s accuracy in predicting the Ivy League Champion. There are now 12 complete seasons of Elo ratings and we will be using all of them to look at how Elo’s success. A successful prediction for Elo happens when the team with the highest Elo rating at the end of the season also wins the Ivy League title. It does not matter if they tied for the title or won uncontested so long as they were one of the winning teams. In the 12 years that have been calculated, the Elo champion was also the Ivy League champion 9 times. This means that, so far, Elo has had a 75% percent success rate. It’s also important to note that in the most recent 5 years, Elo went 5 for 5 in predictions.
One more aspect to keep track of as we continue to calculate Elo is the number of championships each team has accumulated. In this graph, we keep track of the number of Ivy League and Elo Championships each team has accumulated.
Elo seems to have been improving in accuracy as we continue to add more data. We will see how Elo fairs against the real results in our weekly updates.