Early Season Outliers

Posted by on November 11, 2013

It’s probably not a good idea to judge any team based on 40 minutes of basketball. We all know that Duke won’t shoot 76% from two and 62% from three this season. It can be tough to decide if a team is really better/worse than previously thought or if we are simply watching an outlier. To help understand this topic further, I tried answering the question: “What goes right or wrong in outlier games?”

I chose to work backwards to answer the above question. Using a 26,400 game sample from 2009-2013, I used end of season efficiency ratings to predict what each team should have scored on an average day against their opponent also on average day. I was then able to find outliers by finding teams that actually scored at least 10 points per 100 possession more or less than they should have. The result was 4,121 surprisingly “bad” offensive games and 6,756 surprisingly ”good” offensive games.

With these games identified, the final step was to look at what happened to the four factors (shots, turnovers, rebounds, fouls) in these game. Using the same general method as before, I produced the following results:

combined goodbad table

Take a look at the table on the left first. Like most would have expected, it appears unlucky bad shooting is the main reason for a surprisingly bad offensive day. eFG% drops from 49% all the way from to 42%. Unfortunately I don’t have 2P% and 3P% splits, but I would expect 3P% to account for the majority of the eFG% drop off (higher game to game variability). I think it’s pretty intuitive that shooting slumps are more common than offensive rebounding or turnover slumps. The caveat here is that shooting is by far the most important factor. Even if a team was having an “offensive rebounding slump” as long as they hit shots they probably wouldn’t make this list of bad offensive games.

On the right, we see that in good offensive games just about everything goes very right for the offense. FTRate skyrocketing from 37% to 43% is probably misleading, because these teams are likely being fouled when winning at the end of the game. However, these teams did a very good job taking care of the ball. The term hot hand is meant for shooting, but teams can get “hot” (lucky?) taking care of the ball and rebounding too.


Villanova struggled shooting the ball in their season opener vs. a very bad defensive team in Lafayette. However, the Wildcats scored a respectable 1.14 points per possession by taking care of the ball, rebounding, and getting to the foul line. Using what we learned above, I think it’s actually a good thing Villanova shot so poorly. Last year the team struggled mightily with turnovers. It’s actually a sign of improvement that the team was able to score 1.14 points per possession while 5-30 from three. Of course, settling for threes against an athletically inferior Lafayette team is probably not a recipe for long term success. However, simply an early game shooting slump shouldn’t be too big of a cause for concern.

You must be logged in to post a comment.