One of the things I talked about in my previous post was how altering around 6 points per game will get you to around 70% likelihood of covering the spread. I also talked about how each possession has a rough expected value or “ev”.

You can actually get quite detailed in your assessments of what a given possession means in terms of ev for a particular team. For instance the ev of an offensive possession for the Cleveland Cavaliers with LeBron James on the floor is much different than it would be without LeBron James. 

For our purposes we’ll just use very basic league-wide averages to assess the “ev” of a particular play. Here is one example;

This is from a game played on February 26th 2007 with the Miami Heat playing at the New York Knicks.

The betting line “opened” with the Knicks as a 3 point favorite. The closing line had the Knicks favored by 4.5 points. The line move was not due to any injuries or late game information so its fair to say that the line moved 1.5 points due to someone betting a largish amount of money on NYK.

Its fair to say that the bulk of the insider money on this move (the originators) was placed at a line of NYK -3, with the stragglers or those following the bets may have placed wagers at -3.5 and -4. 

Here is the play in question;

One of the things I have talked about before and will delve into further detail about is that you aren’t really going to see too many really blatantly horrible calls. Instead what you’ll see are a collection of rather marginal calls that you could argue could have gone either way. The issue with Donaghy was that it seemed like whenever his bet was in question without exception nearly all of the calls that could have gone either way - went in favor of the side Donaghy’s crew bet on. The above play is a perfect example of this;

To start Miami has the ball with 2 mins left, NYK is up by 3 and the spread in question that NYK will need to cover is also 3 - so its clear that this bet is entirely in jeopardy.

In 2007 on average a team on offense had the following expected point value for possessions;

When not in the bonus 1.034 points

In the bonus  1.084 points. 

Miami specifically on offense had the following numbers;

Not in Bonus 1.02

In the bonus 1.043

And NYK on Defense had the following numbers;

Not in Bonus 1.052

In the bonus 1.11

For simplicity sake lets just go ahead and use the league average expected point number of 1.03 (as MIA was not in the bonus). 

So we have a marginal call that could have been called a shooting foul, or it could have been called a non shooting foul. If it was a shooting foul Miami would have received the 2 points plus a free throw from Shaq who is a career 52% free throw shooter, although he only shot 46% that year. Lets again keep it simple and give him a 50% likelihood of making the free throw and getting an additional point, 13.8% of the remaining 50% of the time that he misses the free throw his team will get the OREB.

So we have the following;

What Donaghy actually called;

Non Shooting foul MIA goes from an ev of 1.03 to 1.08 (they gain .05 points). 

What he could have called;

2 points (for the make)

Plus -

1 point 50% of the time (made free throw) or .5 points

And in the remaining 50% of time for the misses 13.8% will be Orebounded with a basic reset of 1.08 points for a total ev of .5742 points. 

The swing from 1.08 points vs 2.5742 points in one call illustrates how a collection of marginal calls can swing the percentages in your favor. 

The next post will focus more on the actual raw per ref game stats that Donaghy had in a few of his 2007 games.