What does Duke need to do to make the NCAA Tournament? It’s being discussed abundantly, but how can it not be? Duke’s presence on the bubble (using the word “presence” generously here) is a once-in-decades development, and in the broad college basketball picture, the biggest questions of championship week are always bubble-centric. Hard to get worked up about a team trying to climb from a 10-seed to an 8-seed.
With our bracketology model (this morning’s bracket can be found here), we aren’t publishing selection probabilities this year. We never found a good way to calculate out the probabilities of coronavirus cancelations, and we were unsure of their effect (we were also unsure of how to calibrate our error margins with a shorter season than usual, and with an unusually large variance between teams in something as simple as total games played). Were we, though, Duke’s probability of making the tournament would be somewhere between five and ten percent, with a 2.8% likelihood they do so via an automatic bid. That leaves about five percent of simulations in which Duke finds a way without winning the ACC, which implies there is a way.
Now, our model does have some breathing room in its uncertainty margins around the bubble. Saint Mary’s, who appears in just two of the 133 brackets on Bracket Matrix and is our model’s projected fourth team out now that their season is over, makes the field in one of every three simulations of our model. One part of this is that our model is high on Saint Mary’s relative to the industry—most have them further than four places away. The other part of this is that our model’s calibrated fairly broadly—it’s ready to miss by two seed lines anywhere in the field with the same frequency with which it misses by two seed lines in the thick of the field.
With Duke, this means that five percent may be our model’s own caution against declaring them dead. But to check, let’s do the math.
There’s a 2.8% chance Duke wins the ACC Tournament. There’s an 80% chance Duke beats Boston College today. We have those two probabilities.
We don’t have explicit probabilities for the games between—part of the tradeoff we make in amping up our model’s simulations to a high-precision count is that we don’t get detailed winners and losers of each game until Selection Sunday—but we can figure those out in a rough sense. If Duke beats Boston College today, they’ll play Louisville tomorrow as something like a 55% favorite. This means that in roughly 44% of simulations, Duke makes the ACC quarterfinals, where they’d play Florida State. Assuming Duke’s overall rating hasn’t changed much (and our model assumes it’ll stay constant, which is an important thing to note about our model and is responsible for even wider error margins earlier in the year in years in which we do publish probabilities), they’ll be about a 35% underdog. So, making the semifinals happens for Duke in about 15% of simulations. Once into the semifinals, the opponent becomes unknown, but could be Virginia Tech, UNC, Notre Dame, or Wake Forest. In the vast majority of simulations, it’d be Virginia Tech or UNC. Duke would be a small favorite (less than a point) over Virginia Tech. Duke would be a one or two-point underdog against UNC. Knowing that Duke gets this far in roughly 15% of simulations, and that Duke wins the ACC in something like 3% of tournaments, it tracks that if Duke’s slightly less than 50% likely to win the semifinal, making them, say, 7% likely to make the ACC Championship, we can get to that 3% number by pegging Duke as a slight underdog again in the championship—and Duke would be an underdog at this point, should the other side of the bracket produce Virginia, Georgia Tech, or Clemson, the three most likely teams.
In short? Duke makes the championship in something like 7% of simulations. They make the NCAA Tournament in something like 7% of simulations. Make the semifinals, and they may be in Saint Mary’s territory (and our model does expect one in three teams in Saint Mary’s territory to make the field), but they’ll likely be on the wrong side of the bubble, still. Win a semifinal, and they’ll probably cross the threshold. The model isn’t just saying that there’s a chance. It’s got a path.
There’s a lot that can happen. Bid thieves. Other bubble results. The coronavirus. But that’s generally what Duke’s looking at as it starts ACC Tournament play today.
Our movement this morning (no Moving Up or Moving Down because no at-large teams shifted more than one seed line):
Moving In: North Dakota State (auto-bid), Appalachian State (auto-bid)
Holy Mountaineers. App State is in after a head-turning run through the Sun Belt Tournament that featured two overtime victories and four single-digit victories in four days. They currently line up as the 67th overall seed, implying that even other mid-major and low-major conference tournament upsets shouldn’t be enough to push them up out of the 16-seeds.
Out on the plains, South Dakota State lost to Oral Roberts, making North Dakota State the new Summit League favorite. Championship game in that league tonight.
Moving Out: Georgia State (auto-bid), South Dakota State (auto-bid)
These two entered yesterday as the highest 14-seed and the highest 15-seed, which means those that remain have risen. Cleveland State particularly benefited. The Vikings, now more than 50% likely to win out, climbed two seed lines and took over one of the 14’s.
***
Conference Tournament probabilities for leagues in action today. As always, our model relies heavily on KenPom (don’t we all) and our model assumes the coronavirus and other forces will claim no games:
Atlantic Coast Conference
Virginia: 28.1%
Florida State: 25.7%
Georgia Tech: 12.3%
North Carolina: 8.1%
Virginia Tech: 7.5%
Clemson: 5.9%
Syracuse: 3.0%
Duke: 2.8%
Louisville: 2.8%
North Carolina State: 1.9%
Notre Dame: 1.4%
Pittsburgh: 0.5%
Miami: 0.01%
Boston College: 0.00%
Wake Forest: 0.00%
UVA and FSU are the favorites, but only a little better than a 50/50 shot one of them wins it. Lot of teams trying to lock themselves in. ACC’s structure works well in protecting bubble teams from bad losses.
Big West Conference
UC-Santa Barbara: 44.9%
UC-Irvine: 22.3%
UC-Riverside: 19.1%
Cal State-Bakersfield: 9.5%
Hawaii: 2.0%
UC-Davis: 1.3%
Cal State-Fullerton: 0.8%
Long Beach State: 0.2%
Cal State-Northridge: 0.04%
Cal Poly: 0.00%
Play-ins today.
Colonial Athletic Association
Drexel: 61.8%
Elon: 38.2%
Winner’s in, likely eventually on the 15 or 16-line.
Conference USA
Louisiana Tech: 23.8%
North Texas: 21.6%
Western Kentucky: 17.8%
UAB: 14.7%
Marshall: 13.9%
UTEP: 3.0%
Old Dominion: 2.9%
UTSA: 1.5%
Florida Atlantic: 0.5%
Charlotte: 0.2%
Rice: 0.2%FIU: 0.01%Middle Tennessee State: 0.01%
Southern Miss: 0.01%
Our model was unaware of FIU’s cancelation and therefore MTSU’s advancement to the second round.
Horizon League
Cleveland State: 62.3%
Oakland: 37.7%
Championship time here.
Metro Atlantic Athletic Conference
Siena: 26.1%
Monmouth: 22.0%
St. Peter’s: 19.7%
Niagara: 9.4%
Iona: 8.9%
Marist: 7.2%
Rider: 2.7%
Quinnipiac: 2.0%
Manhattan: 1.3%
Fairfield: 0.9%
Rider climbs after winning their opener.
Northeast Conference
Bryant: 74.8%
Mount St. Mary’s: 25.2%
Another championship game tonight. Bryant’s to lose.
Southland Conference
Abilene Christian: 60.8%
Nicholls: 21.8%
Sam Houston State: 14.4%
New Orleans: 1.6%
Northwestern State: 1.0%
Lamar: 0.3%
Southeastern Louisiana: 0.1%
McNeese: 0.04%
Incarnate Word: 0.01%
Houston Baptist: 0.00%
Southland starts its long ladder this evening. Nothing big to see for at least a couple of days.
Summit League
North Dakota State: 53.0%
Oral Roberts: 47.0%
Yet another conference championship, and this one’s about a tossup.
West Coast Conference
Gonzaga: 85.2%
BYU: 14.8%
And in the fifth finale of the day, we have our biggest favorite yet.
