Home > NCAA > David’s NCAA Basketball Tournament Methodologies: SRS and Win Shares

David’s NCAA Basketball Tournament Methodologies: SRS and Win Shares

Final Four: Kentucky, Duke, Kansas, Pittsburgh

Biggest Cinderella:  Morehead State to the Round of 16

When reflecting on my annual futile journey down the road of NCAA March Madness gambling, it seems like the strategies I use each year vary, depending on my success the previous year.  My own intuition always told me that established programs won the final games but were vulnerable in the rounds of 16 and 8.  In a game with equal opponents, I’ll usually place my money on the team with better team defense, and if that factor is a toss-up, the team with a larger front court.  And of course, every single year I am obliged to place the Kentucky Wildcats in at least the Final Four, because they’ve been my team since I was a teenager.  Before we get started, my full bracket is below. 

(Click to Expand)

This year, I’m placing my faith in the all-knowing, all-seeing hands of advanced statistics.  Statistics never lie…right?  Even if I’ve been misinformed on that previous point, there are two advanced metrics that I’ve used to complete my bracket.  Gut feeling and previous loyalty (see: Kentucky Wildcats beating Ohio State in the 3rd round) all play into the picks, but I’m being a little more scientific about my picks this year.  As a result, “four 15 seeds advancing David” has been temporarily banned from the premises.  The two primary metrics I’ve used are:

1. SRS (Simple Rating System)

Taken from the glorious website basketball-reference.com, the SRS rating is a simple value that encapsulates strength of schedule and average win margin.  An SRS value of 0 is considered to be average for NCAA Division I teams.  This year, there are three teams that have distanced themselves from the pack with an SRS rating of 25+ (elite), and those are Duke, Kansas, and Ohio State (highlighted in yellow).

I’m a big fan of this statistic because it truly does factor in not only who you are beating, but by how much.  Teams with high SRS ratings are capable of running up the score on almost any team, and always have more than one “gamebreaker” type player that can go off at any given time.  It is critically important during a long tournament like March Madness to have the elite depth to survive off-games by your big names.

The next pack of teams in the overall SRS rankings has ratings of 21 or lower, a statistically significant difference.  As a result of the importance I’m placing on SRS this year, I think two of those top three overall teams will make the finals.  Duke is my pick coming out of the East/West side and Kansas will make it out of the Southwest/Southeast half. That leaves Ohio State, who I’m predicting will lose by virtue of playing my Kentucky team in the 3rd round.

About 60-70% of the rest of the bracket I filled out using a combination of SRS, strength of schedule, and total team defense ratings.  I factored all three in, and then using my gut feeling chose the winner.  The end result is a bracket that is rooted much deeper in advanced statistics than my brackets of years past.  This is also the reason why I have Washington and Kentucky in the East Final (highlighted in blue above): their SRS rating indicates they should actually both be #2 seeds, instead of the #4 and #7 seeds they currently are.

2. Win Shares

This statistic (again from basketball-reference.com) predicts how many wins a player is directly responsible for over the course of a year, as compared to an exactly average NCAA Division I player.  For example, if Kentucky is 25-5 one year with me playing, and I have a win share rating of 5.0, it would predict that Kentucky would be 20-10 if it replaced me with a player that is exactly average.

Looking at the rankings this year, there are about a dozen players in the 6.7 – 7.8 win shares range (very, very good: each would probably be the MVP of their team).  Above them, there are four players between 8.1 and 8.5 win shares (elite: without them, their team might not even make the tournament).

And all alone in first place is Kenneth Faried from Morehouse State, with 11.8 win shares.

For this reason alone, I’ve got Morehouse State pulling upsets and making it into the round of 16.

Faried is a big senior center who is unlikely to be fazed by the pressure of playing in these tournament games.  He has some robust stat lines in areas that tend to translate well into the tournament (his .644 field goal percentage is ridiculous, as is his 14+ RPG).  He also routinely contributes 2 BPG, 2 SPG and 1 APG to his squad.  The only real downside to his game is his pedestrian 57.8% free throw percentage.  That last number tends to decide games this time of year, so Faried will have to be sharp from the line to fulfill my predictions, although teams in previous years (2008 Memphis comes to mind) have survived without an advantage in that department.

I realize that win shares are a risky statistic to rely on.  Conventional wisdom would indicate that players with high win shares actually play on teams that are more susceptible to losing if the player has an off-day.  The flip-side of this argument is that win shares can actually give you insight into which players are capable of pulling out their Terminator faces and doing serious damage on higher-ranked opponents, and that’s the view I’m taking in the 2011 edition of the tournament.

Win shares for top players also factored into the extended runs I have Purdue (JaJuan Johnson 8.5 win shares) and Arizona (Derrick Williams 8.2 win shares) making this year.

So in summary, this year I’m all about SRS and individual Win Shares to predict the upsets and eventual champions.  I’m confident of success over my fellow bloggers, or at least until Sunday afternoon when I sift through yet another terminally busted bracket.

Quick ending thought: There really needs to be better names for the divisions.  Fans really don’t pay attention to tournament division names the same way they do professional sports conferences.  If the NCAA was smart, they would start selling the naming rights to the four divisions each year, and then donate the entire amount to college scholarships for tournament schools.  The schools already get a big chunk of the monster broadcast revenue brought in from CBS, and this would just add on to that amount.

At the end of the day, I don’t care if I’m watching the Coca Cola division or the Viagra division (actually, the thought of an announcer mentioning that a team is ‘popping out of nowhere’ in that division would be pure comedic gold), and the financial benefits for students would be enormous.  What would those naming rights sell for per year?  5 million?  10 million?  Hard to say, but it would be a lot of coin.

More Strategies from First Off The Bench:

Matt’s NCAA Basketball Tournament Methodology: The Bitter Bracket

Kai’s NCAA Basketball Tournament Methodology: Method? We were supposed to have a method?

Marcus’ NCAA Basketball Tournament Methodology: Battle of the Fiercest Team Names

Advertisements
  1. March 16, 2011 at 2:03 pm

    Very nice! I like the stats that you used to base your picks. How did that Jimmer guy from BYU rank in the Win Shares statistic?

    It will be interesting to see how our brackets do against each other, as mine was picked purely based on irrational emotions, and yours was all rational calculations.

  2. David
    March 16, 2011 at 2:58 pm

    Jimmer was 7th in the nation with a 7.8 win share rating….and that’s with a strong team behind him. If he was playing on Backwater State U, he’d be higher than Faried. Losing Brandon Davies is going to hurt BYU a lot though, and will impair their chances of making a deep run this year.

    Yeah, all four of us have a different approach for sure. I think it’s going to prove that it’s all random luck for who advances the farthest.

  1. March 31, 2011 at 5:18 pm
  2. March 7, 2012 at 7:07 pm
  3. March 14, 2012 at 6:54 pm

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: