Two keys to predicting a major college basketball upset

Stephen F. Austin pulled off a thrilling upset of No. 1 Duke in overtime in November. AP Photo/Gerry Broome

With the Super Bowl in the books, I'd like to welcome you to college basketball season. Has the season been going on since the middle of the NFL regular season? Yes, but don't worry if you're behind. I'm here for you and -- more importantly -- for your bracket. With just six weeks until the first NCAA tournament game tips, here are some links to check out for the statistical story as to where we are today:

No matter how behind you are on your consumption of college hoops, you've heard about the volatility this season. "Madness" seemingly doesn't do it justice: The little number next to a team is beginning to feel more like a curse than a blessing. So let's first take a look at two upsets that shook the sports world to its core.

On Nov. 12, the Purple Aces of Evansville went into Rupp Arena and didn't trail for any of the final 11 minutes against top-ranked Kentucky. Two weeks later, Stephen F. Austin fell behind No. 1 Duke by 15 with the Cameron Crazies in full force and not only bounced back to keep from getting embarrassed, but the Lumberjacks won in overtime in what might be the most exciting finish of the season. Yes, it's true that both of these wins came from teams with sweet nicknames, that the first one came the night of a full moon and the second on a date in which Stephen F. Austin has not lost in 13 seasons, but other than those ironclad trends, what allowed David to beat Goliath?

To be honest, the winning formula didn't prove to be all that complicated: Defend the 3-point line and rebound. In those games, the favored team outrebounded its overmatched opponent by a total of just three boards, a significantly underwhelming number when you consider that those teams combine to average a +12.6 rebound margin this season. Furthermore, the blue bloods converted just 26.5% of their 3-point attempts in these games, way down from their combined rate of 34.5% this season. Is that the solution to the statistical puzzle that is the bracket? Minimize risk by picking teams that have a high rebounding floor and are either very accurate from deep or don't shoot enough triples to bury themselves should they go cold?

This season, there have been 40 instances in which a top-15 team lost to an unranked team. Now, it's important to note that not all of these results are significant upsets in terms of Vegas' point spread, but I have to draw the line somewhere. If you're trying to forecast these sorts of upsets before they happen (see: March Madness), it's worth noting that the losing teams in these upsets have been outrebounded by 2.3 per game. That may not seem like a big number, but when you consider that, for the season, these teams are outrebounding the opposition by 5.0 boards per game, it's a major trend. Here are some teams that may not be on your radar that excel at cleaning the glass:

  • Northern Iowa: 25th

  • Murray State: 20th

  • Stephen F. Austin: 32nd

With the 3-point line moving back nearly 17 inches this offseason, this has been a hot topic, and while the percentage of shots from downtown is down slightly from last season, the accuracy is more the sticking point. In these upset losses, the ranked team is making just 27.5% of its triples, a jarring number and down considerably from 34.2% for the season as a whole. No team is slump-proof from distance, but you can mitigate some level of risk in your bracket by picking the teams that rely less on that shot, thus making an extended slump less impactful. With that in mind, here are some of the big-name programs and their ranking in percentage of shots that come from beyond the 3-point line:

  • Auburn: 51st

  • Maryland: 82nd

  • Illinois: 332nd

  • Kentucky: 348th