Friday Facts: Final Weekend

Well, just another ho-hum no-news Friday in college basketball. Let's get to the Facts:

Purdue, MSU, and OSU are all essentially tied at the top in both raw and adjusted efficiency margins. In fact, it's so close that Purdue leads in adjusted efficiency margin (which is simply adjusted offense minus adjusted defense) but trails both MSU and OSU in conference-only Pythagorean expectancy ("Barthag").

Some might argue that wins and losses matter more than my metrics. By that token, Michigan St. has clinched at least a share of the title, and can clinch outright with a win over the mildly resurgent Badgers in Madison on Sunday:

If the Badgers do pull off the upset, Purdue can grab a share by beating Minnesota and Ohio St. will get a share if they win at Indiana tonight. Here's the slate for the final weekend:

The Myth of Unbalanced Schedules

There's been some kvetching that, given how close the top three are in talent and results, Michigan St. has been gifted the title by virtue of an easy conference schedule.

It is true that MSU has had a favorable schedule: the Facts above show that their projected SOS in conference ranks 13th out of 14th, with only Nebraska having an easier slate. But Ohio State and Purdue are at 11th and 12th. All the contenders had easy schedules, and MSU's wasn't easier enough to talk about a stolen title.

My way of looking at this is to run simulations with a hypothetical double round robin schedule and compare those results to simulations with the actual schedule. I do this every night. Here are the latest results:

These results confirm that MSU's schedule was an advantage in that they were 1.7% more likely get a share of the title in the simulations with the actual schedule than they were in the simulations with the double round robin. But OSU and Purdue were also similarly advantaged. And the advantage in average wins is negligible: .24 over Purdue and .16 over OSU. Assuming those teams finish one game back, they'll have a gripe, but only a small one. They are not going to join the small club of teams who finished a game back and had more than a half-game schedule disadvantage.

The team with a real gripe is Michigan. First, their schedule cut their chances of getting a share of the title by more than half (7.9% to 3.5%). This is because Michigan had a tough schedule and all the other good teams had a favorable schedule. Second, Nebraska's schedule was more than a full win easier than Michigan's. Given that they are locked in a tight battle for the 4th seed and the last double bye at MSG, that is significant. Speaking of which:


At long last, Big Ten teams will be treated next week to the thrill that is playing in the magical arena known as Madison Square Garden. No one can possibly question the wisdom of compressing the conference schedule and subjecting college students to frequent stretches of four games at random times on random days during single-week stretches. If you want an algorithmic take on how the magic will shake out in the end, play around with the T-Rank tourney simulator. Here's the current projection, without trying too hard to get the seeds right:

Four-Bid League?

At this point, the Big Ten has established four locks for the tourney (MSU, OSU, Purdue, Michigan). Three teams—Maryland, Nebraska, and Penn State—still have a chance, but none are looking particularly likely at this point.

Nebraska and Penn State play on Sunday in a pseudo play-in game, of course. Nebraska is the very interesting profile that I've talked about before. If they lose to Penn State my algorithm will pronounce them all but dead barring a deep run at MSG. But by the best metric available—which is WAB (or SOR)—they'll still be very much on the bubble even if they lose Sunday and lose their first game in the tournament. Teamcast projects they'd be ranked 52nd in WAB in that scenario—outside the field, but certainly on the bubble. One win at MSG (really avoiding a bad loss) moves them all the way up to 41st, which is definitely tourney-worthy.

Ultimately, I don't have much confidence that my algorithms are correctly placing where Nebraska comes out in any of these scenarios. But I do know that they need just one more win to have definitively amassed a tourney-worthy resumé.

I'll hopefully explore this and more in a pre-MSG post next week.