VegasWatch has been invariably aggregating an ingenious calculation of rating teams using KenPom efficiency ratings adjusted to the average Vegas number.  This concept allows for a very flexible data set, which one can use to determine a variety of results in relation to a team’s performance or expected performance.

I’ve decided to convene this comparable data on a per conference basis, using numbers gathered from BBState (raw statistics not adjusted to schedule) to calculate a team’s conference efficiency adjusted to relative tempo, and formulating the aforementioned with the average conference spread for each team.  What this can do is measure Vegas precision for each conference.  By Vegas I mean linesmaker precision, the offshore industry really rules the market nowadays, but in common vernacular, Vegas is a term that is denoted for aspects regarding the general spread.

Here is the aggregated data set from the Big East with a graph to demonstrate the varying magnitudes for the represented teams (Updated through February 22):



The data is sorted based on the average conference spread per team in descending order. The first column represents their average spread accumulated throughout the conference season, meaning only Big East games were used in the data.  The next column labeled “Conference Efficiency Margin” are the possession stats extracted from BBState measuring average efficiency margins per conference game (A positive efficiency is displayed as it would be a Vegas line, which is used to discern an easier relationship to compare the numbers, therefore a team showing a positive number in the table actually has a negative efficiency that would betray its true line in retrospect).   Finally the differential was formulated through a simple difference, which resulted in an average differential for the conference. A negative differential number for a team would suggest a team has been underlined throughout the conference season, a positive number would suggest the opposite.  For the Big East, the average differential was 1.4343, which is an absurd level of accuracy on so many levels.  This basically is a reaffirmation of Vegas acuity, it is really stunning.  What this data suggests also, is getting the best number is important, as what has been substantially proven through the two point line move research.  Just under 1.5 accumulated points separates the spread for Big East conference games from linesmaking perfection.

Again this data is very flexible.  The numbers could be used to convene a line for a game, make numerous conference future predictions, and even expected record against the spread for every team.

Here is the data from the remaining five major conferences, all data displayed in a graph with the same magnitude of eight for consistency (at length I will do some of the more prominent Mid-Majors, perhaps the margin of error grows by degree as the conference ratings decrease). 

The graphs portray a margin of error between the line and efficiency for the top team in each conference to be very accurate, with one minor exception, Kentucky, which approaches a differential of -2.  This would indicate the Wildcats being underlined during SEC play.

(There was a recalcitrant bug in OpenOffice Calc that jutted the X-Axis for some graphs despite my better efforts to alter the situation for the better)

Continue reading here