In the point guard study I performed last week (http://basketball-statistics.com/howtheshootingabilitiesofpointguardsaffectoffenses.html), a few questions arose.  Perhaps the most important and potentially the most problematic was the question of whether or not three point percentage is useful simply because it raises a player’s overall true shot percentage (and therefore efficiency).  If the players with the highest three point percentages also had the highest true shot percentages, it’s not much of a surprise that three point percentage leads to a positive impact on the team’s Offensive Rating.

To try to answer this question, I factored in true shot percentage to the original data.  I broke the point guards down into four groups: those with above average 3PT% and above average TS%, those with above average 3PT% and below average TS%, those with below average 3PT% and above average TS%, and those that were below average in both.  Theoretically, if three-point percentage is valuable independently of true shot percentage, then point guards with a high 3PT% should have very positive results on their team’s Offensive Ratings regardless of their true shot percentages.  Of course, some complications do arise when doing the analysis this way, and I will get into that later.  For now, here are the results:

Before I get into analyzing those tables, I should get into the complications.  The biggest problem is still the fact that a high 3PT% correlates with a high TS%.  Let’s look at an example.  If Player A is in the high 3PT%, high TS% group, it is likely he’ll have a higher TS% than Player B, who is in the low 3PT%, high TS% group.  Even though they are both in high TS% groups, Player A is a great three-point shooter, and that happens to be a very efficient shot.  To see what kinds of problems those complications cause, let’s take a look at the averages:

The latter of those two tables is what we care about the most.  As we can see, the problem is much more of a concern among the below average TS% groups.  In those two, having a point guard with a high three-point percentage makes all the difference.  In the two above average TS% groups, the difference is only about two percent, which is significant but not very substantial.  Let’s keep this in mind as we go back to analyzing the original three tables…

When we look at the table for “ALL,” the original study seems to be invalidated.  The Offensive Ratings follow the true shot percentages – the highest average true shot percentage is in the above average 3PT%, above average TS% category, and that’s where the highest Offensive Rating is.  The two second-place categories match up, and so on.  Basically, all that really matters is the true shot percentage of your point guard, not his three-point percentage.  What happens when we look at the home/away splits?

This is where things get interesting.  First, let’s look at the “Away” results.  Here, it is pretty clear that true shot percentage is all that matters.  The two high TS% categories fare much better than the two low ones.  In fact, even though the high 3PT%, low TS% category has an average TS% nearly five points higher than the low 3PT%, low TS% category, the difference in Offensive Ratings is insignificant.  It’s almost as if the ability of your point guard to shoot three pointers is a detriment on the road!

At home, things swing completely.  As far as I’ve noticed, there are two results that indicate that three-point shooting for point guards at home is especially important.  First, there is the difference in the two above average TS% categories.  Despite having an average difference in TS% of just about two percent, the difference in Offensive Ratings is nearly five points per 100 possessions.  Secondly, look at the difference between the above average 3PT%, below average TS% group and the below average 3PT%, above average TS% group.  Despite the former having a true shot percentage over six points lower than the latter, the Offensive Ratings are close.   The two observations I made in this paragraph don’t really make any sense, unless of course there’s something special about three-point shooting for point guards when you’re at home.

I’ll leave it to you to figure out reasons why this may be so.  I’ve got a couple of ideas, but they are all just guesses.  There are also a lot of other observations that can be made about all of the data, and certainly a lot of spin-off studies could be performed.

Copyright © 2009 Basketball-Statistics.com