Tuesday, December 11, 2012

Underdogs

It's time to get your defibrillators warmed up for Saturday night.  For the first time this season, the Wildcats are legitimate underdogs, at least that's how things appear.  Basketball games are not the best tools for figuring out which of two teams is better.  They are simply too short and therefore severely affected by matters of chance.  Over 50 to 80 possessions plus overtime when needed, one team will prevail, but sometimes it is not the best team.  If we played Florida 1,000 times, we would know for sure which team is better.  Instead, we get them for 40 minutes, or maybe 45 or 50.

I want to start by looking at this matchup the way I have done for other teams in the past.  Then, I want to take things to a whole new level of nerd. In the side by side comparison of stats, it's a mixed bag.  Florida has a better FG%, but we shoot the three better.  We get more points per shot, but the two teams are equal in points per possession.  Florida has a decisive edge on the defensive end.  We are better on the defensive glass; they get more offensive boards.  We are bigger.  They are more experienced.  We have more depth.  You can see why there is a lot of hype about this game.



What comparisons like these don't account for is strength of schedule and past competition.  Florida has put up these stats against some good teams, like Marquette, Wisconsin, and Florida State.  Comparatively, our schedule has been weaker.

To look at things a little more closely, I decided to simulate the game. I've been hesitant to get into the prediction business because there is no better way to lose favor with folks than to make predictions that are not very accurate.  Still, not many people read what I write, so what am I really risking?  A basketball game is actually very easy to simulate.  In any given possession, there are only a few things that can happen, and the probabilities of each can be established because prior games can be used as guides.  The only slightly tricky part is that I adjusted the likelihood of shots falling relative to the defensive proficiency of each team as measured by Ken Pomeroy.

Here's what my computer says would happen if the Cats and Gators squared off 10,000 times.  It says that Arizona would win 40.2% of those games. It also says that roughly 1 out of 40 of those games will go into overtime.  The average score  would be 68.6 - 65.2 giving Florida about a 3.4 point margin.  Yes, we are underdogs.  I should note that KenPom only gives us a 37% chance of victory and predicts a loss of 67-63.  He uses different methods, and his technique, unlike mine, incorporates home court advantage.  I have not.  If I did, it would probably give us a 50/50 chance of winning.

Below are the simulated point distributions for Arizona and Florida as well as the distribution of point margin.  Note that a huge range of outcomes is possible.  If we had an amazing game, and Florida had a terrible day shooting the ball, of course we could win by 40. There is about a 1 in 1000 chance of that happening.  The reverse could happen as well. The chance of Florida winning by 40 or more?  About 1 in 160.  As I take my first foray into college hoops prognostication, keep in mind that a huge range of outcomes is possible, but a three to four point loss and a game in the mid to high 60s is most likely.  Let's hope I'm wrong about the first part.


No comments:

Post a Comment