Are we #5 or #15?

The MSU basketball team, as ranked by various methods on January 28, 2010:

The humans like us quite a bit more than the computers (or, rather, the equations calculated by the computers) do.

Why is that?

First, MSU is probably a little overranked in the human polls, even using human-poll standards.  Texas should clearly be ahead of us (one fewer loss; beat us head to head).  Probably Duke, too (4-1 vs. AP top 25; MSU is 2-2).  Essentially, there's a big pile-up after the top four--9 of the 11 teams between #5 and #15 have 3 losses--and the voters are giving MSU the benefit of the doubt because they've lost least recently.  Plus MSU started with the advantage of being #2 in the preseason polls.

Second, MSU is 4-1 in games decided by 4 points or fewer.  That means they don't look as good in the purely points-based computer rating systems (KenPom and PREDICTOR) as their binary win-loss record would indicate.  Only beating Iowa by 7 at home hurt quite a bit in the points-based systems, too.  I think we dropped 5-6 spots in the KenPom ratings after that game.

Of course, MSU is also outside the top 10 in the point-margin-ignoring ELO_CHESS system.  The merged Sagarin rankings only have MSU as 2-3 vs. top 50 teams, the wins coming against Wisconsin and Gonzaga.  The underwhelming performance of the Big Ten in nonconference play isn't helping MSU here.

LVS's links post from last night included a piece by Ken Pomeroy in which Pomeroy examined his own rating system to figure out why it's so down on Kentucky--which was #1 in the polls at that point, prior to losing at South Carolina, but just 9th in the KenPom ratings (#11 now).  Here's what he had to say about margin of victory (which is the main driver of his ratings system):

While a team plays to win the game, its chances of winning are much greater when it leads by 20 with five minutes to go than if the game is tied at that point. There is a significant incentive to building a comfortable lead rather than just trying to stay a point ahead of the opposition all game long. You never know when your opponent is going to go all Chandler Parsons on you. For this reason teams capable of building big leads typically build them, and past results indicate that those teams are headed for good things in the future.

He went on to note that Kentucky had won quite a few close games this season (4 by 5 points or fewer, including 2 against sub-KenPom-100 teams).  The math says that tendency was bound to catch up to Kentucky--and, lo and behold, it did Tuesday night.

Pomeroy pointed to one team he thinks wasn't necessarily subject to the laws of basketball rating system mathematics: the 2006 Gonzaga team that could coast on defense because of Adam Morrison's unstoppability on offense.  Fellow Prospectuser John Gasaway added the 2007 edition of the Florida Gators as another team that--due to its potent mix of NBA-level talent, experience, and cohesiveness--could apparently impose its will and escape the gravitational pull of its Pythagorean calculations.

So does any of this have any bearing on MSU's current disparate positions in the human-intuition-based and math-based rankings?  I'd say probably not to a large degree.  It certainly doesn't feel like MSU has been "coasting" through games and then flipping a switch to pull away at the end.  It feels a lot more like they've struggled with some fairly fundamental areas of play (ball-handing, in particular), found ways to keep games close, and been poised/fortunate enough to pull out several of them late.

There is one X-factor that we may be able to argue is an analytically-significant variable that the computer rating systems don't know about.  He wears the #1 jersey for MSU.  Kalin Lucas has now established a pretty long track record of coming up big in the final 5 minutes of most close games the Spartans find themselves in.  Off the top of my head:

• Home to Wisconsin last season (jumper to break tie; 3-pointer to go up 5).
• At Illinois (twisting layup to break tie and spark run).
  
• Kansas in the Sweet 16 (and-one vs. Collins to break tie).
• Home to Gonzaga this season (layup to go up 3).
• At Minnesota (pull-up 3-pointer to break tie).
• At Michigan (jumper to take lead).

Maybe I'm forgetting some close games in which Lucas didn't make the play(s) needed to win the game, but I think most neutral observers would agree there aren't many players a college coach would rather have with the ball in his hands and the clock running down in a one-possession game.

At the end of the day, MSU's position in the rankings is going to have a pretty marginal impact on its Big Ten and NCAA fortunes.  We're dealing with pretty fine margins here.  The difference between the #7 and #15 teams in the Sagarin PREDICTOR ratings is only 1.37 points.  The Spartans are winning games right now, which is the main thing we care about as fans.  And the numbers say they're in striking distance of being among the nation's top-tier teams, in a year in which it doesn't appear there's a truly elite team on par with, say, last year's North Carolina squad.

Having a proven late-game playmaker doesn't hurt, either.

P.S. Purdue barely escapes at home against Wisconsin.  That's a good result for us, I think.  We're now up 3 games in the loss column on everyone else in the league.  Had Wisconsin won, they'd be looking pretty dangerous.  KenPom doesn't have them favored to lose a game from here on out (50/50 at Minnesota).  If they'd won tonight and you assume they beat us at their place next week, suddenly they're only a game back.  As things are, you still have to view them as a solid threat.  They've played remarkably well in Jon Leuer's absence.  Once he's back, they will be extremely tough to beat.  (It doesn't look like he'll be back for the MSU game, by the way.)