Game Analysis

Navy (-2.5)  17   Army  14

I’ll start with the headline. The Army-Navy game has gone under the total 16 consecutive seasons and all games between the 3 military schools are 51-15-3 Under going back many years.

As I’ve mentioned for years, teams that run the option offense practice against it every day and are best at defending it. I used to play the unders in the option team versus option team games but that trend has become well known and the totals are now about 10 points lower than they’d be otherwise. In this case, a points model would predict a total of 43.6 points in this game and this game opened with a total of 33.5 and that number is even lower now.

The fact that the under trend in these games is well known has not stopped these games from going under even with drastically lowered totals. The last 6 games between military academies have all had totals of 41 points or less, which is much lower than would normally be projected, and all 6 of those went under the total by an average of 13.6 points.

The method I decided to use to project points in this game while still respecting the trend is to use the regression equation projecting points scored in the Army-Navy game as a function of the predicted total of a points based model for each game for the last 16 years (all unders).

There is some correlation between the total scoring (points scored and points allowed) of these teams for the season entering this annual season-ending game and how many points are scored and these two teams are relatively lower scoring than the majority of matchups over the last 16 years. In this case, a compensated points model would project 43.6 total points and the regression equation, using the last 16 games between these teams as data points, would project 30.9 total points scored. Keep in mind that I only used the games during the last 16 years with each game going under and undoubtedly the projection would be higher using the same method if I went back back further.

Thus, even if the trend is still valid, the value is basically gone. However, I’ll respect that trend and project 31 total points, which is not enough value to make the under a profitable play even if the pattern of lower scoring than expected games in this series continues as it has the last 16 years.