The NBA season starts tomorrow and the Golden State Warriors will begin their title defense at home against the New Orleans Pelicans.  The Warriors are currently 9.5-point favorites and the total is set at 214.5 at Pinnacle.  These lines would suggest Golden State winning 112-102.5 (assuming half-points were possible).  The question is, how accurate are these totals?  Can you use these to help handicap your fantasy team? Or who to start on DraftKings?

I looked at the closing lines for every regular season and postseason NBA game last season and did the same calculation.  I then compared the suggested* team total to the actual points scored by both teams.  I separated them by home and visiting teams to see if there was a bias towards one or the other.

*I use the word suggested because the actual team total posted at Pinnacle might have been slightly different than just calculating using the closing lines.

Below are the graphs of the difference between the actual points scored and the suggested team total at Pinnacle:

Visitor TT Diff

Home TT Diff

There are clearly some outliers, but you can see the pattern of a normal bell curve in both graphs.  Calculating the mean for each data set shows that each is around zero (zero meaning the points scored equals the team total):

 Visiting TeamsHome Teams
Standard Deviation10.610.4

A standard deviation of over 10 points highlights the variance that we see in the NBA.  While a majority (68.1%) of games are expected to fall within +/- 10 points of the projected total, that means that nearly one third of all NBA games will not fall in this expected range.

While a closing line at a sharp book should be considered the best line for a game, it still can’t account for the variance in “shots falling” in any given night.  Whether handicapping games, teams, or individual players, make sure you know there is a wide range of results to expect in the NBA.