If you’re a college hoops junkie like myself, you spend all year watching basketball for hours nearly every night from mid-November all the way up until Selection Sunday in the middle of March. Watching every team at least twice. Scouring the country for that sleeper team in one of the mid-major conferences that’s going to make the difference.
What difference, you ask? The difference in your bracket pool of course. Whether it’s a big pot of cash or just good old-fashioned pride on the line, you want to be the smartest guy in the room.
Maybe you spend all of Sunday night through Thursday morning poring over every match-up. Fine-tuning here, tweaking there until that noon tip arrives. Or maybe you’re one of those first instinct people just going off the top of the dome.
Me? I’m a fine-tuner. I have to analyze every detail until I feel comfortable. And by the time that first tip-off happens, you can’t convince me that there’s a single flaw in my bracket. It’s perfect every year. Dream big. Shoot for the stars. Bet on yourself, man. “Warren Buffett’s gonna be so mad when he has to give you a billion dollars,” I think to myself.
And guess what? By about 2:45 P.M. EST on Thursday last year, I had already lost one of my teams in the national championship. That’s right, I had the Iowa State Cyclones in my final game. Absolutely loved their path to the championship. Paste this 14-seed UAB, take care of SMU or UCLA in the second round – didn’t really matter, neither of them were near the Cyclones’ level – and then beat Gonzaga in the Sweet 16 because it’s Gonzaga in the Sweet 16. Basically a triple bye to the Elite 8. Sure, Duke was talented, but ISU would beat them with experience and all of a sudden I’m one win away.
But before you knew it the Cyclones were getting pounded on the boards by that pesky Blazer team that was decent on the glass, but certainly not the group of Dennis Rodmans Iowa State made them out to be. Just like that my bracket was worth more as a fire-starter than a retirement plan. So long private island. The private jet will have to wait another year.
There’s a lesson to be learned here, though, and you can’t put a price tag on this lesson. Much like a big man trying to guard the post, you’ve got to do your work early. So that’s what I did.
As my good friend George Santayana once said, “Those who ignore history are bound to repeat it.” If I just would’ve looked into the history of it all, I would’ve seen a couple of big flaws with my belief that Iowa State was a good bet to make the Final Four.
About that history. I wanted to know what the statistics said about the teams that did reach that third and final weekend. What common traits do they share? What areas do the most successful teams do well? Was VCU just a hidden gem in 2011 that we should’ve seen coming?
The answer to that last question is no. Of all the stats I gathered from KenPom.com dating back to 2002, VCU is the biggest outlier. They are the worst defensive team (out of 56) to make it to the Final Four and they were only better offensively than five other teams. They just got extraordinarily hot at the right time. But it was the right year for a team like that to make a run (we’ll get to that later).
I’m sure at this point you’d just like me to give you the damn data already so you can take all of your friends’ money next month. I’m not much for small talk myself. Shall we?
I’ll get to the overall offensive and defensive efficiency stats later (SPOILER ALERT: Final Four teams typically rank pretty high in both categories), but in this section I want to address more of the specifics. Once again, all these stats are courtesy of KenPom.
Something to keep in mind here: for practically all of the 46 areas I evaluated, you’ll find a team ranked in the top five in the country and you’ll also find at least one ranked in the high-200s or 300s. There’s more than one way to skin a cat. I’m just looking for the biggest trends.
One last quick note: there will be a lot of references to national rankings. For a reference point there are 351 teams currently, up from 327 in 2002, which is the first year of data.
The three-point shot seems to become a bigger part of basketball every year. The best team in the NBA basically lives and dies by it (they live most of the time). The leader in the clubhouse for player of the year in college – Buddy Hield – puts up 8.5 of them per game. The guy nipping at his heels down the stretch – Denzel Valentine – throws up 7.6 shots from the outside on an average day.
But while the attempts nationwide continue to increase every year, most of the teams you see playing in April are below the national average in the rate that they shoot from outside the arc compared to their total field goal attempts. Put simply: the best teams typically don’t shoot that many threes. The average rank of Final Four teams is 204th. Only 2 of the 56 have ranked in the top-10 percentile.
While teams typically shoot above average from outside, it’s not an overwhelming number. In fact, it’s much more important that teams complete a high percentage of their shots on the inside. Besides overall offensive efficiency, 2P% is the biggest indicator of success on that end, along with Effective FG%.
That’s not to say that 3P% doesn’t matter, though. There’s only been six teams that have ranked in the 200s in that category and none that have ranked in the 300s. Teams have to be a threat, but the data shows that it’s not a good sign if a team relies on the deep ball.
I’m saving my final analysis on teams for another day, but I’ll tell you right now who this section is a bad omen for: the #1 KenPom rated Villanova Wildcats. They’re 21st in 3PA/FGA and they’re 263rd in 3P%. Last year, they had a nearly identical 22nd attempt ranking, except they hit them at a much higher rate and they still lost in the second round to an 8 seed.
Oklahoma – led by Hield – launches a ton of threes as well, but they also hit the highest percentage in the country. But in their four lowest marks percentage-wise this year, three of them have resulted in losses. Much like that Nova team of a year ago, one bad shooting game and this Sooner team could be toast in the first weekend.
Duke, Indiana, and Wichita State all put them up pretty frequently, as well.
One of the coaches I used to work under always used to say you can typically boil the game down to two stats (OTHER THAN POINTS, WISEACRES): turnovers and the following category in this list, rebounds. His belief was that if you win the turnover and rebounding battles, you’ll win 80% of the time. Not that I needed the data to support this claim, but it turns out he’s pretty smart.
The average Final Four team ranks 78th in offensive turnover percentage. Defensively, teams are right around average (167th) in forcing them, but that still results in a net gain.
It’s not rocket science. Basketball is about possessions, especially in, say, a close tournament game.
Who might this be an issue for of the contenders? Maryland. They rank 261 in turning it over and 283 in turning their opponents over. That issue has come up in three of their five losses this year. They had 22 against North Carolina, 15 against Minnesota and 12 against Michigan. 12 may not seem like much, but it was a low possession game and the Wolverines only turned it over 8 times in a one score game. That can be the difference between a first weekend exit and a deep run.
Once again, another element of a game that affects the number of opportunities you have to score. Not surprisingly, the best of the best have an average rank of 66 in offensive rebounding percentage and 100 on the defensive end.
I don’t have much else to add. Good teams rebound.
There’s actually quite a few teams at the top that are pretty average rebounders, but given that there’s a little more historical emphasis on offensive rebounding, Virginia and Iowa State are the ones that stick out. Virginia at least cleans it up on defense. ISU is bad at one end and average on the other.
Defending the Perimeter
While it’s not essential to shoot the three well, history shows that it’s pretty important to defend the three well.
The worst team defending the three in the past fourteen years was actually last year’s Wisconsin team, coming in at 309th in the nation. The average is 87. It almost cost the Badgers in the second round, too. They ran in to a hot shooting Oregon team, who was an 8 seed.
Who struggles in this area? Funny enough, Oregon does. It’s a big issue for North Carolina as well.
Defending the Interior
Guess what? The battle of the paint is more important defensively, as well. Outside of the overall efficiency numbers and effective FG percentages, it’s the lowest average of any other area at 48. In fact, there have only been two teams below the national average that have made the Final Four. They were both in 2011: VCU and Butler. I’ll get in to it later, but that was the most wide open year that the data covers.
That seems like a very telling stat, so who among the top 25 or so contenders doesn’t meet the criteria? Well, Iowa State ranks 159th and Baylor is 193rd. More bad news for the Cyclones. Baylor’s on the fringe of even being a contender, but their only losses outside of the brutal Big 12 schedule were on the road to Oregon and Texas A&M. So good news, bad news for Baylor fans: you might be a contender, but this stat does not bode well for your contender status.
BREAKING NEWS: Fouling is bad. Teams who don’t send their opponents to the line and allow them to get an average percentage of points on free throws fare better in basketball games.
With that said, I don’t think this has that big of an impact in the long run, but it’s certainly a factor.
Teams that have some fouling issues: Kansas, Michigan State, West Virginia (par for the course for a pressing team), Kentucky, and Wichita State.
This one only has data through the 2006-07 season, but I found the data over that nine years kind of surprising. You always hear the talking heads talking about teams having enough depth and if they have a lack of depth, it’s a problem. The average Final Four team over those nine years ranks 244th in bench minutes.
There have been 15 teams in those nine years that have ranked in the 300s in bench minutes while there have been only three teams who ranked in the double-digits. One of those teams was Kentucky last year, which technically doesn’t count because most teams don’t have five McDonald’s All-Americans coming off the bench.
So who’s playing too many guys? Michigan State, West Virginia, North Carolina, Wichita State, Purdue, Indiana, Texas A&M, Texas, Valparaiso, and Dayton. Uh, that’s a lot of teams! It’s hard for me to imagine all of those teams missing the Final Four, just from a sheer numbers standpoint. I mean, Izzo’s team is in there. The Final Four is his second home.
I’m not sure what you do with teams like Duke and Iowa State this year, though. At this point, they’re both going about six-deep. They both fit in with playing a shorter number of minutes with their bench, but they’re also one more injury from playing five guys.
This one I found interesting. The data is only for the last six years, but on average teams are about average on offense at a 177 pace and below average at a mean of 239. Any team that is ranked in the double-digits on one end is significantly slower on the other. There’s only 1 out of 24 teams that have been in the double-digits for defensive pace – Wisconsin 2013-14 – and they were 342nd in offensive tempo.
It’s obviously limited by the lack of years, but it seems kind of significant, no? It makes sense if you think about how games tend to slow down in the tournament. Teams that are comfortable winning games that way are more likely to succeed.
Take last year’s national champion, Duke, for example. They played a slightly above average pace all year. They played three games all regular season that were well-below average pace and they won all three – included in those were wins over Louisville and Virginia on the road.
They proved that they could win games that way and four of their eight games in the postseason were played well-below an average pace. They won all four.
Who doesn’t fit that mold this year? Xavier. That’s about it.
How Champions Become Champions
All of this analysis has been on the Final Four teams, but what do the actual national champions do differently? Nothing really, they just do everything a little bit better, on average. They turn the ball over even less, shoot less threes, rebound even more, and defend better. The bench minutes are about the same and champions have actually played slightly faster, but they still fit the same general mold.
The most dramatic differences between the mere semifinalists and those that win it all lies on the defensive end. “Defense wins championships” isn’t just a cliche every coach wants you to believe after all. The average rank of a national champion in effective FG% on D is 22nd. The worst ranking? 70th from Duke last year.
Which brings me to another important point: all these numbers so far have come from the post-tournament data. Overall efficiency numbers are available for before the tourney, but individual pieces would take a lot more work. Most teams’ numbers don’t change all that dramatically during the five or six games they played in the bracket.
I mention that because Duke’s overall defensive efficiency did change dramatically during their run last year. Going into the tournament they were 57th and they came out of it 12th after winning the national championship. They were an above average defense that became a great one when it mattered most. Recency matters and the Blue Devils held three of their last five opponents before the tournament under 55 points.
Shot blockers have proven to be a big help, as well. The average for a champion is 46th, compared to 87th for all Final Four teams.
Now we get to the overall numbers: offensive and defensive efficiency. Think of every other area I’ve touched on as an exam score. These numbers are the final grade.
Average OffEff for Final Four teams: 20
Average OffEff for national champions: 13
Average DeffEff for Final Four teams: 24
Average DefEff for national champions: 19
Those numbers are pre-tournament numbers.
It’s no surprise that those numbers are low, but I was more interested in how many teams were elite in both categories. In order to do this, I wanted to see how many teams ranked in the top-35 heading into the tournament in both categories. I chose the 35 number because it’s the top 10% of the country.
The answer? 68% (38 of 56) ranked 35 or better on both ends. There wasn’t a major difference in champions, interestingly enough. 71% (10 of 14) fit the bill and the last two winners – Duke and UConn – did not.
So, about that other 32%. Does it pay to be elite at offense and just okay at defense or vice versa? Not really. 84% (47 of 56) were elite offensively and 80% (45 of 56) were elite defensively.
But you have to be great at one of them. Only 2 of the 56 teams have made the final weekend without meeting either criteria. Guess who? Our 2011 friends: VCU and Butler.
Another important thing to note is that no team that’s made the Final Four in these last 14 years has been below average on one end. So while teams don’t have to be great on both sides of the ball, they have to at least be passable. The data tells you that if you were thinking you just might outsmart everybody and put Notre Dame in Houston this year because they have the best offense, don’t. They’re 232nd defensively.
How Number One Teams Fare
Speaking of teams that are the best in a category, how have the data-driven #1 teams heading into the tournament done?
The most efficient offenses heading into the tournament have made 3 of the last 14 Final Fours, as have the most efficient defenses. Two of those offenses have won the title – ’09 North Carolina and ’08 Kansas – and the ’13 Louisville team is the only number one defense to win it all.
The number one rated team overall on KenPom using his Pythagorean rating has reached the last weekend six times and won it all three times.
Here’s the breakdown of all 14 of the #1s in KenPom’s ratings:
2002: Duke – lost in Sweet 16
2003: Kentucky – lost in Elite Eight
2004: Duke – lost in Final Four
2005: Illinois – lost in National Championship
2006: Duke – lost in Sweet 16
2007: North Carolina – lost in Elite Eight
2008: Kansas – won National Championship
2009: Memphis (2 seed) – lost in Sweet 16
2010: Kansas – lost in Round of 32
2011: Ohio State – lost in Sweet 16
2012: Kentucky – won National Championship
2013: Louisville – won National Championship
2014: Arizona – lost in Elite Eight
2015: Kentucky – lost in Final Four
Speaking of the Pythagorean rating, the average Final Four team is ranked at 10.5 and the winner is right around a 7. There have only been three winners outside of the top-7 – ’03 Syracuse, ’11 UConn, and ’14 UConn.
Is It Actually Wide Open This Year?
History suggests that if you’re a chalk bracket player, you’re gonna have a bad time this season.
For one, maybe the biggest narrative of this collective college basketball season has been all of the upsets happening this year and top-5 teams falling left and right. I saw a stat on ESPN the other day that we were about to break the record this year for AP top-5 ranked teams losing. The current record-holder is the 1979-80 season. The 1980 Final Four was composed of a 2 seed, 5 seed, 6 seed, and 8 seed.
Another common narrative this season is the belief that there are truly no great teams. Numbers back that up. Going back to the Pythagorean rating, there are zero teams this year with a rating of at least 0.95. Compare that to last year’s field, which had five teams that met that standard.
The average is two teams per year that meet that 0.95 mark. So I wanted to see the difference in seed totals of teams that reached the Final Four during years with a few 0.95 teams and those with one or zero. Listed below is each year and in parentheses is the number of 0.95 teams and following is the seed total of the four semifinalists. Years with one or zero teams are in bold.
2002 (2) – 1, 1, 2, 5 = 9 seed total
2003 (1) – 1, 2, 3, 3 = 9 seed total
2004 (2) – 1, 2, 2, 3 = 8 seed total
2005 (2) – 1, 1, 4, 5 = 11 seed total
2006 (0) – 2, 3, 4, 11 = 20 seed total
2007 (2) – 1, 1, 2, 2 = 6 seed total
2008 (3) – 1, 1, 1, 1 = 4 seed total
2009 (1) – 1, 1, 2, 3 = 7 seed total
2010 (2) – 1, 2, 5, 5 = 13 seed total
2011 (1) – 3, 4, 8, 11 = 26 seed total
2012 (3) – 1, 2, 2, 4 = 9 seed total
2013 (2) – 1, 4, 4, 9 = 18 seed total
2014 (2) – 1, 2, 7, 8 = 18 seed total
2015 (5) – 1, 1, 1, 7 = 10 seed total
You can see that the two biggest seed totals happened in years like this, which was also the only two years that a 1 seed didn’t make it. At the same time, ’03 and ’09 indicate that it’s not a guarantee that it’s going to be absolute chaos in the brackets. 2009 was actually a pretty chalky year overall. Every team in the Elite Eight was at least a 3 seed and there was only two teams outside of the top-4 seeds that were in the Sweet 16.
While signs seem to point to more of an open year, there are still some ways to narrow down contenders using the Pythagorean rating. 77% of teams that make the Final Four have a rating of at least 0.90. 93% have had a rating of at least 0.87. The four teams that didn’t make the 0.87 mark were ’06 George Mason, ’11 VCU, ’11 Butler, and ’13 Wichita State.
If you were to narrow this year’s field down to just the 0.87 teams, you’d have 20 to choose from. There are 22, but two of them – Louisville and SMU – are ineligible, which is another factor this year. The loss of Louisville removes another major contender from the field.
20 is on the lower side of number of teams at that 0.87 mark. The only other two years with 2 or less 0.95 teams and 20 or less 0.87 teams: 2006 and 2011. Once again, the two craziest years we’ve had in the past 14.
So, What Does This All Mean?
It means that it’s probably not in your best interest to wager heavily on too many aspects of this year’s tournament. There are a lot of indicators that things are going to get a little wonky in March.
At the same time, it’s impossible not to get in the mix with at least a couple of bracket pools. My best advice would be to rely on teams that fit the mold laid out in the “Common Traits” section and to pray (if you’re in to that kind of thing).
What teams most closely align with those “Common Traits” is a subject I’ll address in the coming weeks.