Ben from Dear Old UVa is once again back to statistically analyze the NCAA Tournament for us.

Last time, I told you about my dorky little model.

This time, we’ll take a look at the second half of the finals: UConn and Michigan St.

First, UConn:

To me, the amazing part about UConn’s season is how much they turned it around to make the final four.  Look at that graph.  I guess I could’ve put a trend line on it, but it clearly would’ve been downward-sloping.

Then, you get to the tourney and there’s a huge discontinuity.  It’s clear why most people undervalued them, but the model actually appears to overvalue their regular season, giving them only one loss (to Louisville).

Michigan State, on the other hand, looks a lot more like Villanova:

The model predicted a few more losses than the Spartans suffered and it clearly picked them to lose to Louisville.  It tells you that Tom Izzo does a great job of preparing his team for the tournament when they really outperform expectations this much.

Fittingly, then, the model picks Michigan State to lose to UConn by six points (OK, actually 6.1 points).

Now, I’m not a gambling man, but I wanted to see how this stacked up against Vegas and Kenpom.  The table below shows how it stacks up.

So, if you want my advice*, give the points on North Carolina and UConn. 

I feel like UConn’s going to win a close game, by more than five points.

Anyhow, enjoy the games everyone!

* Note: Like I said before, I’m not a gambling man.  So if you follow my advice and lose, well, that’s your fault, not mine.  Don’t sue me.

rtmsf (3998 Posts)

This entry was posted on Friday, April 3rd, 2009 at 3:28 pm by rtmsf and is filed under 2009 ncaa tournament. Tagged: 2009 final four, connecticut, michigan st, statistical analysis, unc, villanova. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.