Five more seasons have been added to the Ivy League Football Elo ratings. The 1958 to 1962 seasons are shown below and all Elo ratings calculated are shown at the bottom of the page.
There have been a few updates to our Elo ratings. The formulas are still the same as the ones explained in the introduction article; however, some aspects of the graph have changed. The most obvious change is in the y-axis, which we expanded from 1300-1700 to 1200-1800 because of Dartmouth’s dominance and Brown’s failures. Another note is that in the 1961 season there was a tie for the title between Columbia and Harvard. They also happened to finish around the same Elo rating, which makes it hard to see both of their championship markers on the graph.
Now that there is some more data for the Elo ratings to work with we can see how well Elo ratings stacks up against the real Ivy League Champions. The Elo ratings will also show us which team was the most dominant over this period.
First let’s look at the Big Red, who did not see too much success during this period. Cornell peaked at a rating of 1561 in ‘58, but their rating fell to 1358 in ‘61 before leveling out around the mid 1400s. These ratings show that Cornell was in the lower half of the Ivy League. They were usually able to beat some of the worse teams and occasionally pull an upset, but were unable to maintain the success needed to get a better Elo rating.
In this period, there is one team that is the clear Elo champion: Dartmouth. The Big Green finished at the top of the ratings in all but one year when they settled for the second-best rating. Dartmouth also won 2 Ivy championships in this period. They were able to achieve the highest rating of any team yet at a score of 1722 at the end of their ’62 season. The reason for such a high rating is their continued success over the course of these seasons since Elo rewards consistency.
However, the Elo ratings were only able to correctly predict the Ivy League Champion about 60% of the time. The ratings are not perfect and could be slightly skewed for a multitude of reasons. Football teams’ abilities can change quickly if a key player gets injured, coaches change or the team plays well against a stronger rival. Elo cannot account for extra factors and takes several games to adjust to large changes. If the ratings were more reactive then they would adjust to every upset or blowout quickly, but consistency would not be rewarded as much. The key to Elo is to balance the team’s history and the team’s current performance in order to see their true strength.
Comments