A (sort of) scientific way to fill out your March Madness bracket
03/18/2008
Paulsen Home / Sports Channel / Bullz-Eye Home
It’s bracket time!
All over the country, millions of workers are blowing off their responsibilities to pour over whatever stats they can find -- seeds, records, rankings, strength of schedule, common opponents, RPI -- to fill out their March Madness bracket. Eventually, they’ll come up with one that they’re comfortable with, pony up their $20 and hand it in. On Tuesday and Wednesday, they’ll try to get some work done so that it’s not that obvious that they’re missing work on Thursday and/or Friday afternoon.
A typical fan will fill out multiple brackets so that by the time the Final Four rolls around, they’ll probably still be alive in at least one pool. I don’t have a problem with people filling out multiple brackets, but you’re only allowed to talk about one when boasting to your buddies. It’s not cool to brag about how you picked three different upsets when you picked them in three different pools. Give a monkey a pen and enough empty brackets and he could have done the same thing.
Last year, I used a series of stats (which I referred to as the “Six Cat”) to pick winners in the tighter contests. The results were so-so. I did pick the overall winner (Florida) along with one other Final Four team (Georgetown), but once the Hoyas lost to Ohio State (who should have lost to Xavier in the second round, damnit!) in the semifinal, I was finally knocked out of my pool.
This year, I’m tweaking my system a little bit to focus only on those stats that show a strong trend:
Sagarin ratings
Jeff Sagarin has been issuing his computer ratings for years, and I’ve found that they are a good predictor of upcoming games. What qualifies as good? I started keeping track during last year’s tourney, and teams who had the better predictor score by two or more points were 42-7 (86%). Not bad.
Seed advantage
Patrick Reilly over at BostonSportsHub.com has analyzed the last 16 years of NCAA tournament games and has found that a team with a 4+ seed advantage wins at least 75% of the matchups.
Points Per Shot
Reilly ran the numbers for several team statistics, and most of the trends weren’t overwhelming. However, the team with an advantage (2% or more) in regular season field goal shooting was 66-41 (62%) in matchups of similar seeds. I am going to take this a step further and use Points Per Shot (PPS), which divides a team’s total points by the total number of field goal and free throw attempts. PPS takes into account total shooting, so if a team is terrific from three-point range or is great at getting to the free throw line, it will be reflected in this statistic. The average PPS for all the teams in the tournament is 0.953, and when comparing teams, we’ll use this as a reference.
Conference tournament fatigue
Reilly also examined how fatigue from Championship Week might influence the outcome of first- and second-round games. He found that teams seeded #3-#6 were 14%-15% less likely to survive the first two rounds if they played three or more conference tournament games.
Using these four trends, along with a little common sense, I hope to put together a bracket that will once again keep me competitive all the way through the Final Four.
Let’s get started....
FIRST ROUND
Using Sagarin ratings, we can pick the winners of 25 of the 32 games since there are teams in those matchups that are at least two-point Sagarin favorites. The 4+ seed advantage in three of the 6/11 matchups eliminates three of the seven remaining games, leaving us with four games to analyze.
#8 UNLV vs. #9 Kent State
#7 Gonzaga vs. #10 Davidson
#8 Mississippi State vs. #9 Oregon
#7 Miami vs. #10 Saint Mary’s
None of the teams in these four games fit the conference tournament fatigue criteria, so we’ll only look at shooting percentages. In the first matchup, Kent State is 2.7% better than UNLV, so we’ll go with the Golden Flashes. The Gonzaga/Davidson game is a matchup of two very good shooting teams. Gonzaga has a 1.2% advantage in FG%, but Davidson is 4.1% better in PPS, so we’ll go with Davidson. (It doesn’t hurt that they are riding a 22-game winning streak or that the game is in Raleigh, NC just a three-hour drive from the Davidson campus. Meanwhile, Gonzaga has to fly cross-country.)
The Mississippi State/Oregon matchup is an interesting one. The Ducks hold a 2.4% advantage in FG% while their PPS is 7.8% better. I’m a little worried about taking Oregon over Mississippi State since the game is in Little Rock, but I’m going to trust these shooting numbers. Besides, the Pac-10 was the tougher conference this season.
In the final matchup, St. Mary’s holds a 2.3% advantage in FG% and a 2.6% advantage in PPS. Let’s go with Gaels.
Here is a full list of my first-round picks: North Carolina, Indiana, Notre Dame, Washington State, Oklahoma, Louisville, Butler, Tennessee, Kansas, Kent State, Clemson, Vanderbilt, USC, Wisconsin, Davidson, Georgetown, Memphis, Oregon, Michigan State, Pittsburgh, Marquette, Stanford, St. Mary’s, Texas, UCLA, Texas A&M, Drake, UConn, Purdue, Xavier, West Virginia and Duke.
SECOND ROUND
Sagarin ratings yield solid picks for 12 of the 16 second-round games, leaving four matchups to examine more closely.
#4 Washington State vs. #5 Notre Dame
#4 Pittsburgh vs. #5 Michigan State
#3 Stanford vs. #6 Marquette
#4 Connecticut vs. #5 Drake
Neither Washington State nor Notre Dame went deep into their conference tournament, so both teams should be relatively fresh for Thursday’s game. The shooting edge goes to the Cougars, who hold a 1.9% advantage in FG% and a 1.5% advantage in PPS, so we’ll go with them.
Pittsburgh fits the mold of a team that may suffer from conference tourney fatigue. After knocking off three ranked teams -- #13 Louisville, #24 Marquette and #8 Georgetown -- to win the Big East Tournament, they have to hop on a plane to Denver to play the Thursday/Saturday set of games. A potential matchup with Michigan State looks to favor the Spartans, but I’m going with the Panthers, who are suddenly one of the hottest teams in the country. You don’t beat three ranked teams in three days unless you’re for real.
The Stanford/Marquette matchup is tight across the board, though Marquette holds a small (1.8%) advantage in PPS. Stanford played three conference tournament games but only has to travel to Anaheim for the first-round games. It’s a much longer trip for Marquette and I think that offsets Stanford’s potential fatigue. For that reason, I’m going to go with the Cardinal.
Drake holds a 7.0% advantage in PPS so they would seem to have the advantage over Connecticut. Three conference tournament games might cause a little fatigue, but two of the games were blowouts and the Bulldogs don’t play until Friday, so let’s assume they’ll get their rest. I’ll take Drake in the upset.
Here’s a full list of my second round-picks: North Carolina, Washington State, Louisville, Tennessee, Kansas, Clemson, Wisconsin, Georgetown, Memphis, Pittsburgh, Stanford, Texas, UCLA, Drake, Xavier and Duke.
SWEET SIXTEEN
Sagarin ratings give us solid picks in six of the eight Sweet Sixteen games, but I’m a little worried about picking Wisconsin over Georgetown even though they have a 2.2-point advantage. The Hoyas are a full 8.5% better in PPS and they have the defense to match the Badgers’. So I’m going against the plan and picking G’town.
Here are the two matchups that are too close to call:
#2 Tennessee vs. #3 Louisville
#2 Texas vs. #3 Stanford
Tennessee holds a slight advantage in FG% (0.8%) and PPS (1.8%), but Louisville is a solid team from a tough conference. The game is in Charlotte, which is closer to Knoxville than it is to Louisville, and my guess is that the Vols will have more fans there. Their loss in the conference tourney led to a #2 seed, so they’re probably angry at the perceived snub. This should be a great game, but I think Tennessee will avoid the upset.
Stanford/Texas looks like a great matchup on paper. The Longhorns hold a 4.6% advantage in PPS and the game is in Houston, so I think Texas will come out on top.
Here are my Sweet Sixteen picks: North Carolina, Tennessee, Kansas, Georgetown, Memphis, Texas, UCLA and Duke.
Those are all #1 and #2 seeds, so there are no real reaches there. I do think that Pitt would have a good shot to upset Memphis and that Xavier could topple Duke, but I don’t feel strongly about either upset to buck the Sagarin odds. Over the last 16 years, 64% of Final Four teams were #1 or #2 seeds, so it generally pays to go with the chalk.
ELITE EIGHT
Sagarin ratings yield three solid picks in the Elite Eight, but I think that Memphis will go down before the Final Four due to their style of play. They want to push the tempo, and it leads to some questionable shots down the stretch. They’re going to have a tough time beating Texas in Houston, especially considering the Longhorns’ athleticism and terrific guard play. The Derrick Rose/D.J. Augustin matchup would be a beauty.
The other tight game is #1 UCLA vs. #2 Duke.
This matchup is tight across the board, but UCLA has found ways to win the tight ones all season and I don’t think that they’ll allow Duke to beat them with the long ball. Kevin Love will give the Blue Devils fits down low, so I think the Bruins will punch a ticket to the Final Four.
Here are my Elite Eight picks: North Carolina, Kansas, Texas and UCLA.
FINAL FOUR
Kansas/North Carolina would be a terrific matchup. The Jayhawks hold a slight Sagarin advantage (1.8), and a substantial FG% (2.3%) and PPS (4.7%) advantage. Kansas always seems to disappoint in the tourney, but that reputation has a lot to do with Roy Williams’ tenure there. Ty Lawson is the key. I don’t think he’s 100% and the Kansas guards should be able to contain him.
The Final Four is in San Antonio, which is going to make it more difficult for UCLA to prevail. But the Final Four has a notoriously “corporate” crowd, so I don’t think the virtual home court advantage will be as strong for Texas as it was in the regionals. The Bruins hold a 2.5% advantage in FG% and a 1.2% advantage in PPS. They also have a terrific defense that should be able to offset Texas’ guard play.
Final Four picks: Kansas and UCLA.
CHAMPIONSHIP GAME
Kansas vs. UCLA
The Sagarin ratings (Kansas +2.6) say that this is a Kansas win. The Jayhawks also hold a distinct advantage in FG% (2.7%) and PPS (4.9%). Kansas has the size and athleticism to limit Kevin Love, and the speed on the perimeter to control Darren Collison. UCLA will keep it close, but I think Kansas will win this game by 8-10 points.
As you can see, this system didn’t pick a bunch of upsets. Though they happen each and every year, they are so tough to predict that you’re better off just going with the favorites. It’s rare for a #1 or #2 seed to miss the Sweet Sixteen and many of them go on to the Elite Eight, making high seeds tough to pick against in the early rounds.
So, now that Jeff Sagarin and I have done most of the research for you, fill out your bracket and hand it in. Feel free to change up a few of the picks to honor your alma mater or to pay tribute to the teams with the hottest cheerleaders. Or change everything up. That’s the beauty of March Madness -- we all have our own ideas on how it’s going to play out.
Now get back to work. You have a lot to do before Thursday rolls around.
Send questions and comments to jpaulsen@bullz-eye.com.