Wednesday, December 18, 2019

About the Akey Bracket Composite (ABC)

Hi! Welcome to the Akey Brackets Blog.

One of the main focuses of this blog is regularly publishing the Akey Bracket Composite, or "ABC." The ABC is a method of bracketology for Men's NCAA Basketball. It compiles seven metrics the NCAA selection committee uses. These metrics are divided into two categories: results-based metrics, which rank teams based on their record and strength of wins, choosing the "most deserving" teams; and the predictive metrics, which rank teams based on efficiency and other advanced stats, choosing the "best" teams. These seven metrics are:

Results-Based
-ESPN SOR
-KPI
-RPI
-NCAA NET

Predictive
-KenPom
-Sagarin
-ESPN BPI

The NCAA actually uses all of these to make decisions, and even explored using a composite of six of these metrics before creating the NET ranking system instead. However, the initial tournament bracket using this system in 2019 did not follow the NET rankings very closely. Would a composite of these rankings be more useful, or accurate?

The Composite
The ABC has gone through the past few years of data and found how closely each of the above seven ratings correlates to the field the NCAA selection committee chooses. (NET has been used for only one year, so there is currently a small sample size for this metric. The years where RPI was relied upon more by the committee will be slowly discarded.) Using this data, each metric is given a historical multiplier based on how well they predict the bracket. Each team's seven rankings are multiplied by their corresponding multiplier and summed, creating an ABC Value.

These ABC Values result in a numerical reflection on the committee's customary decisions. By ranking teams with these values, we can create a consistent, less subjective method of bracketology.

Thanks for visiting!