Grady’s Picks
As march starts to wind down, one of the best times of the year in the sporting world is upon us: March Madness. This year I have decided to take a different approach to predicting my picks to try to see though the future madness that is ahead. In the past I would base my picks just off team resume’s, randomness, and spreads of the game. This year, I am going to base my bracket from one thing: historical seed success in the tournament.
The first and most important thing about my bracket is the champion. 1 seeds have won the tournament 25 out of the last 39 times since 1985. That’s good enough odds for me to fill in the two 1 seeds that I believe will fare the best in the tournament. I have Duke taking over Florida in the national championship.
Analyzing more in the final four, all one seeds have only made it one time, in 2008. On average the most common result for one seeds in the final four is 2 teams, leaving room for two more. The other most common teams to make the final four are the 2, 3, and 4 seeds, to no surprise. I have picked the other teams from the field that I like the most that are not one seeds, being Iowa State and Alabama, to make the final four alongside Duke and Auburn.
Analyzing the round of 64, 1 seeds almost never lose, so by nature my champion Duke, runner up Florida, Houston, and Auburn move on. About the same can be said with 2 vs 15 seed matchups, so Tennessee, Alabama, St John’s, and Michigan State will be moving on. 9 seeds find themselves beating 8 seeds about 51% of the time, so I have picked my two favorite 9 seeds to move on, Creighton over Louisville, and Baylor over Mississippi State. Leaving Gonzaga over Georgia and UConn over Oklahoma. Moving on, 10 seeds beat 7 seeds about 38.7% of the time so I have decided to pick my favorite 10 seed, New Mexico to upset Marquette, leaving Kansas over Arkansas, Saint Mary’s over Vanderbilt, and UCLA over Utah State. Continuing the trend of upsets 39% of 11 seeds upset 6 seeds, about 1.5 wins for each tournament. To get a little riskier I picked my two favorite 11 seeds, VCU to beat BYU and SDSU/UNC to beat Ole Miss, leaving Missouri over Drake and Illinois over TEX/XAV. 12 seeds upset 5 seeds 35% of the time, I have picked my two favorite 12 seeds to upset the 5 seeds staying a little riskier, UC San Diego over Michigan and Colorado St over Michigan, leaving Clemson over McNeese and Oregon over Liberty. 5 seeds lose to 13 seeds about 21.5% so I picked one 13 seed that I liked the best to move on, which is Yale over Texas A&M, leaving Maryland, Arizona, and Purdue to make it through. Finally, to end the round of 64, 14 seeds upset 3 seeds a mere 15.3% of the time. While its more than possible, it’s too risky for my bracket, so Kentucky, Iowa State, Texas Tech, and Wisconsin going through.
Here is a graphic from the NCAA to back up some of the Percentages I used to justify how many specific seeds would be moving onto the round of 32.
These seed success rates have allowed me to pick the best combination of teams that I believe will be good while following the best percentages of upsets to happen. Having both the round of 64 and the final four set already this makes the rest of the bracket must easier. Attacking the round elite 8 the same way that I did for the final four it makes my decisions much easier for the sweet 16 and the round of 32. I did consider certain seed success in the round of 32 and sweet 16, but at the rest of the bracket was also done accordingly to fit my final four picks.
I have never done a bracket this way, but I wanted to test it out and see if there could be a serious advantage to trying to follow the most likely outcomes. March is always Madness, and it seems kind of redundant to follow a certain method like this as your whole bracket can blow up at any given moment if one of the championship teams go down early, but that’s why we all love this time of march and why no bracket has ever been done before. I plan on doing two other brackets not following this method and just using “gut feelings” and “ball knowledge”. After the tournament is over, I will compare the results to see if this is a useful method to follow in the future.
