In a few hours the first-ever 16-team Big East Tournament will tip off, bringing new meaning to the term “heavy underdog.” You thought Georgia was an inspiration in the SEC tournament last year, winning four games in four tornado-blighted days? That was nothing! From this point forward any team in the bottom half of the Big East that wants an NCAA tournament bid (assuming they don’t already have an at-large pocketed), will have to win five games in five days. When that happens it’ll officially be time to pull Al Michaels into the sport for a “Do you believe in miracles?” 2.0. He’s had years to come up with one.
Speaking of pulling folks into action, it’s a special occasion so I dragged Ken Pomeroy kicking and screaming out of semi-retirement to provide me with his patented log5-based preview of the tournament.
What’s a log5-based preview? Take it away, K-man:
The log5 method was developed by Bill James more than 25 years ago. It’s based on sound theory and provides a way to derive a team’s chance of winning a particular game using each team’s winning percentage. In the case of college hoops, we’re forced to use a schedule-adjusted version of winning percentage, one I derive from the adjusted offensive and defensive efficiencies I compute.
When applied to tournament play, we can get a good estimate of each team’s chances to snag its conference’s automatic bid. It’s only an estimate, though, subject to the assumptions that my model of adjusted winning percentages is reasonable and that a team’s personnel has not drastically changed late in the season. With all that said, the numbers produced are reasonable and give us an idea of how likely a team’s path to the promised land may be.
In a log5 table, all figures listed indicate a team’s chances of getting to a particular round, with the rightmost column listing a team’s probability of winning it all.
And so for example in this instance we find that DePaul‘s path to the promised land is not very likely:
Seed Qtrs Semis Final Champ
3 Connecticut 100.0 72.3 41.3 27.1
1 Louisville 100.0 84.4 54.0 24.8
2 Pitt 100.0 61.9 33.0 20.7
7 West Virginia 76.8 33.2 15.4 8.6
4 Villanova 100.0 53.9 24.0 8.3
6 Syracuse 82.5 25.5 8.9 3.9
5 Marquette 55.8 26.2 10.6 3.2
12 Georgetown 42.2 19.6 7.8 2.3
10 Notre Dame 22.5 4.8 1.2 0.4
8 Providence 56.0 8.9 2.1 0.3
9 Cincinnati 42.3 6.5 1.4 0.2
11 Seton Hall 14.4 1.7 0.2 0.1
13 St. John’s 2.0 0.3 0.0 0.0
14 S. Florida 3.1 0.4 0.0 0.0
15 Rutgers 0.8 0.1 0.0 0.0
16 DePaul 1.7 0.2 0.0 0.0
The number that stood out to me is pretty well buried in there: Ken’s winning percentages really like Syracuse in their second-round game against the winner of Seton Hall vs. South Florida.
Alas, I digress. Let’s stick with the obvious headline: our eyes say there’s a big three in the Big East and the numbers agree emphatically. There is a great yawing chasm between the third most likely outcome (Pitt winning it all) and the fourth most likely outcome (West Virginia winning it all).
Within that big three we see a very interesting and, I think, instructive dynamic. Yell all you want about “settling it on the floor,” but the truth is that settling it on the floor entails seeding. And seeding is huge. Look at Connecticut. They’re actually less likely than Louisville to make the semis and the finals, yet the Huskies are more likely to win it all than are the Cardinals. How can this be?
Because Jim Calhoun‘s team has a much tougher path to the finals. Louisville gets to play the winner of the game between Providence and either Cincinnati or DePaul, while UConn will most likely be up against Syracuse. No offense to the Friars, but that’s a huge benefit in the Cardinals’ favor.
Speaking of benefits, Pitt certainly didn’t get any when they were matched with (likely) the winner of Notre Dame vs. West Virginia in the quarterfinals. It’s tough to imagine a team that’s a lock for an NCAA one-seed getting a more difficult pairing in the quarterfinals of their own conference tournament. Recent accolades notwithstanding, Jamie Dixon‘s team better be ready to play from the get-go.