March Madness and Analytics

Like many other outlets, I am doing a March Madness related post. I have decided to take an analytics slant to this. I often hear people say that they can do better than analytics can. I wanted to very loosely and enjoyably put this to the test. We could think of it as my version of Dwight vs. the computer from the Office.

It should be a fun post and I won’t take this too seriously because there is a lot of randomness in sports that occurs. But it is a fun look at how these three methods shake out.

The Set Up

So I have created 3 different brackets. The first one is purely my picks without having looked at any analytic line specially for this (see picture below). Though I am posting this while the first games are being played, I have not looked at any scores. I only know who won the First Four games (at the top of the bracket pics).

The second bracket I completed was based solely on the website fiverthirtyeight.com (see picture below). They do a whole slew of different metrics and analytics, including the NCAA tournament. I picked the teams that they said would be most likely to win for every single game. There was one game that was 50/50 and I picked the team that they said would be most likely to win it all between those two choices as the tie break (Illinois/Baylor match up).

For the third bracket, I randomly picked teams using the =rand() command in Excel to give two random values (again, pictured below). The team assigned to the higher value was the team that I would pick. This bracket turned out pretty wacky and I am excited to see how it turns out

Caveats

Here are a few things that might influence the different brackets I have completed. First a bit about me. I enjoy basketball and have played pickup basketball for many years. I do not have a lot of years of organized basketball. I typically follow college and NBA basketball from a distance though and have watch just a few games from my alma mater, BYU this season. In other words, I’m not a terribly deep SME on the topic. As you can see, I put a bit of my heart into the bracket when I picked BYU over Texas in the second round.

Second, I did not look at any game progress and completed my picks before looking at fiverthirtyeight.com for that bracket. As previously mentioned, I did see who won the First Four games but I have not seen any other game progress or scores before making my picks. As I followed news about my BYU Cougars I did see a couple sports pundit opinions before making my picks, but nothing I saw was extensive or detailed. Just a few predictions about upsets.

Third, I came up with my picks by just choosing the teams that I thought would win. I usually go to something like FiveThirtyEight and look at a few teams schedules late in the year. But I did none of that to create my bracket. So, there was less thought than normal put into this bracket. That may actually help me based on my previous bracket performance (which has not been good).

Scoring

I will use a basic scoring scheme for these. I will multiply the see by the round for every team that moves on and add up the score at the end. At the end of the tournament, I will see which bracket had the greatest number of points.

Final Thoughts

It’s interesting to see the differences between these brackets. The final four for my picks and the FiveThirtyEight are very similar. The only difference is that I picked Houston instead of Illinois. I also picked a bunch of upsets that could be big point values. The random bracket is really fun. It features a VCU, Bama, Winthrop, and Cleveland State final four. It should be interesting to watch!

As always, thanks for reading!

–Brandon
Subscribe to the blog below!