dizzysoft software

I am a huge college basketball fan so naturally this is my favorite time of the year. This is probably because I went to the University of Kansas- a school who’s only coach with a losing record is James Naismith. My love of all things Jayhawk Basketball makes it hard to fill out my brackets in any objective way- somehow I always end up with Kansas winning the championship. That means I only predicted the national champion once since I was in Junior High School (the last time the Jayhawks won the National Championship).

This year I had to come up with a better method for filling out my brackets. Everyone has their own method- from considering RPI to team mascots- but I wanted to do something different. This year I used Googlefight to pick the winner. As I did, I learned a little about keywords and SEO.

Basically Googlefight compares two terms and determines a winner based upon the number of results that term comes up in a search. So, when I put in “North Carolina” vs. “Radford” naturally North Carolina came up with more results and thereby won. My theory was, the more people talking about a team the more likely that team would win. This would consider all important factors in an objective way- if a smaller school had a lot of buzz around it, it could trump another school, for example.

There came a problem when I got to some terms. Some were so broad they prevailed over others that the picks were rather silly. For example, “American” vs. “Villanova” resulted in American winning (actually, they almost did). Another problem cam up when I entered proper names for schools that never use them- everyone calls “Connecticut”, “UConn” for example. I knew I had to be more specific.

Since I was really wanting to measure discussion around a school’s basketball team I decided to narrow my focus (lengthen my keyword tail) to include “basketball” after the name of the school. I was afraid that using abbreviations for team names wouldn’t be accurate enough so I used the formal name of the school, excluding terms like “universtiy.” This resulted in preferring schools named on geographies rather than private schools because “Texas basketball” would consider any basketball related ideas within Texas. I think this was good (we will see) because schools recruit in their geography so if there is a lot of interest in basketball in that state, that school tended to win.

Some of the results were interesting. For instance, when I entered my beloved Jayhawks against my Dad’s alma mater- NDSU, the Jayhawks merely edged by. That’s exactly how the game felt! At the same time, no matter how I entered “American Univeristy” it always kept winning. This is because this term is so broad- and had I not made an exception to my method here, American would have won the tournament.

As of right now (Saturday afternoon) my method has picked 23 out of 33 games correctly and predicted some surprising upsets: Siena, Arizona, Cleveland State, Dayton, Texas A&M, Maryland, and Michigan.

Follow me on @dizzySEO to see how my brackets conclude. I will use the hash: #marchmadness.