On March 27, 1939 the first NCAA Tournament was born and with it came the social phenomena of March Madness. Originally only having an 8 team format, the tournament has evolved into a 68 team campaign for the Championship. Since its inception 35 different teams have won the “Big Dance” with UCLA leading the pack with 11 titles followed by Kentucky at 8. UCLA won 10 of its 11 tittles behind the reigns of Coach John Wooden between 1964 and 1975.
Though no one knows for sure when filling out brackets for March Madness first became popular, the rise of the Internet in recent years has exponentially increased the amount of brackets completed and the popularity of this competition. This has led to networks like ESPN and Yahoo to have a nationwide competition awarding people for the best brackets.
Most notable of these competitions, is this year’s Warren Buffet and Quicken Loans Billion-Dollar Challenge. The challenge is simple: select a perfect bracket, win a Billion-Dollars. Now as simple as this sounds, the odds are against you.
You have a 0.000000000000000005421% chance of picking a perfect bracket if you randomly do every pick. That’s approximately less than One in Quintillion chance. You have a better chance of winning the lottery twice before you select that perfect bracket. But lets say you consider yourself an expert in college basketball and have a 90% chance of picking a game correctly. You would still have less than a one percent chance of correctly picking every game.
Still feel like your bracket is “The One”? Good, me too.
After looking through the results from every tournament three key findings were found and should be in consideration as you make your picks.
- A No. 16 Seed team has never beaten a No. 1 Seed
- A No. 1 seed has a pretty good chance of making it to the Final Four
- Since they began seeding teams in 1979, a 1 Seed has made to the Final Four in every year with the exception of 1980, 2006, and 2011.
- A No. 11 Seed is the lowest Seed to ever make the Final Four
- Everyone loves a Cinderella story, but before you pick NC State to make a Final Four run remember this
Bleacher Report’s C.J. Moore came up with 10 Traits that almost all of the Previous Champions had in common. Here they are:
- Head Coach has made it to the Elite Eight in the past
- Has either won its regular season conference title or conference tournament.
- Ranks in the top 20 in kenpom.com’s adjusted offensive efficiency ratings.
- Ranks in the top 20 in kenpom.com’s adjusted defensive efficiency ratings
- Shoots better than 37% from beyond the arc
- Has at least three double-digit scorers
- Has a front-court scorer who averages better than 12 points per game and will eventually get drafted in the first round of the NBA draft
- Rebound better than 37 percent of their misses on the offensive end
- Holds opponents to less than 45% shooting inside the arc
- Has a defensive free-throw rate better than 31%. (Free-throw rate is the number of free throws attempted divided by field-goal attempts)
Using data analytics is it possible to create a Predictive Formula using Data Analytics?
Well the quick answer is no. Though it is very possible, as of right now no one has cracked the code. A lot of people are able to pick who will win or who will make the Final Four, but to come up with a correct prediction for every game is nearly impossible. There is just too much randomness.
I decided to take a whack at this and try to predict the results of 2014 NCAA Tournament. With some help, I collected the data starting from the 2006-07 season up to the most recent. Using each teams statistics during the regular season I ran a regression analysis to get each team down to one single number.
Below are the basic statistics from all of the teams:
After running the regression I found that a few statistics had to be rejected because they represented no correlation between how well a team will do in the Tournament. Some of those variables are: Rebound Margin, Assists Per Game, Strength of Schedule, and Points per Possession.
After removing these variables, my model only was still only able to represent 42% of the variation in Tournament results. This truly shows how random March Madness is. Hear is the predicted results:
If you have any suggestions about any article ideas or sports data that you would like me to analyze, feel free to contact me at email@example.com .
Special thanks to CamRondo (@CRJustice) and Megan (@BabichMegan) for helping with some of the data entry!