Earlier today, I wrote an article entitled “Using Advanced Metrics to Separate Contenders From Pretenders in This Year’s Tournament Field“. The article is based on: the thoughts of someone I trade sports debate emails with; Luke Winn; Ken Pomeroy; and, most specifically, the adjusted defensive (AdjD) and offensive (AdjO) efficiency metrics available at KenPom.com.
I think that there are a lot of relevant and illuminating tournament-relevant nuggets of information in the article, so I encourage you to read it; but if I were to boil the conclusion of the article into one sentence it would be this: the teams most likely advance furthest in the NCAA Tournament are those with low AdjO and AdjD numbers.
So, because I am obsessed with March Madness and the NCAA Tournament, I decided to put what I learned this afternoon into use and create an alternative bracket to compete with my official bracket predictions.
In this post, I’ll explain the simple bracket predicting methodology that I used, based on the KenPom AdjO and AdjD metrics, as well as show you the results of the bracket.
Before I explain the methodology or show you the bracket, let me reiterate something that mentioned in the “Using Advanced Metrics…” article linked above:
I am not switching my official bracket picks.
Over 100,000 people have viewed my original bracket prediction article, which creates an objectively picked bracket based on my 7-metric system chosen for 2011, and that bracket remains my bracket for 2011. I’ll live and die with it.
I am publishing this bracket simply to provide additional guidance for MSF readers and as a test for next year, because I am already excited to incorporate AdjO and AdjD stats into my bracket prediction system for 2012.
With that said, let me explain my very simple methodology for creating this bracket:
- First, I went to KenPom.com, my new favorite college basketball website, to download data for the 2011 season.
- With this data, I isolated the AdjO and AdjD metrics, specifically the national rankings of each team.
- I used the following formula to calculate a value: AVERAGE(AdjO + (AdjD*.95)) – this provides the average of the AdjO and AjdD numbers while weighing defense 5% higher than offense. (Why the weighted average? Well, to use a cliche, because defense wins championships. What can I say – I’m a Bob Knight disciple. All else equal, give me the team that’s more efficient on defense. I wanted the picks to slightly reflect this bias.)
- If you want my spreadsheet, here you go.
- I went game-by-game all the way through to the championship advancing the team the lowest number.
And here are the picks, which I added to the MSF Bracket Challenge (where you should add your picks!):
If you can’t read this, click here to view larger image in a new window.
A couple observations:
- Not surprisingly, all four #1 seeds made the Final 4. In fact, in all but the Southeast Region, the Regional Semifinal games are comprised of the top 4 seeds. This is highly unlikely to happen.
- The Southeast region does feature a good bit of turmoil before the Elite 8. This method predicts 12th-seeded Utah State and 13th-seeded Belmont to win their opening round games and meet up in the Round of 32. #11 Gonzaga and #10 Michigan State also advance to the next round before bowing out.
Now, you may be thinking to yourself: what a boring bracket! And for the most part, you’d be right. Other than some first round madness in the Southeast, chalk pretty much holds to form, and we know that is not going to happen.
BUT…
As much as we like to talk about upsets and Cinderellas come tournament time, you are still usually best served by picking the teams that have proven themselves over the full season to move on. Predicting where the upsets will come can be like finding needles in a haystack. You can try, but it is a very inexact science based, oftentimes, more in luck than actual prognosticating skill.
And here’s the thing: this system does predict five double-digit seeds to make the Round of 32. That’s actually a pretty good number. Four of them are just clustered in one region, which makes the bracket seem too chalky.
Anyway, who knows how this bracket will play out. Like I said above, my official picks remain unchanged. I am just intrigued by this methodology, and I wanted to share it with you all in my quest to make you as informed as possible when it comes to filling out your brackets. We’ll see how this bracket does.
What do you think? Do you like this bracket predicting methodology better than my “official” bracket prediction system for 2011? How would you meld the two?
About Jerod Morris
