How Aggregate NBA Stats Change Through Season?

Getting me interested in any random NBA statistic is easier than convincing a baby to eat some candies… and that’s exactly how this post started. Andres Alvarez made a following note on Wages of Wins Journal:

I would love to see a chart of how stats on aggregate look if we chart them over a season.

And I was hooked…
Especially when we often hear/read that “shooting picks up as season goes on” or anything similar.
Is it true? How various statistics do change through a course of a season?

To investigate I used box scores from ESPN.com and one line for each team named “TOTALS” for every game in the last 3 regular seasons. Then I divided season into 7-day sections but not by a date but literally by “days of games in a row” so first seven were Week1, another 7 days were Week2 etc (this way All-Star weekend and any random breaks didn’t affect sample size of each week, outside of last and week all have 92+ teams’ lines). Finally I calculated averages for each category and here are some results [keep in mind that in each graph seasons are marked with the same colour so red is 2008-09, green is 2009/10 and blue means 2010-11]:

So assists and turnovers have indeed a sizeable period of time at the beginning of each season when teams either experiment with playbook or simply can’t figure out their rotations.

I can’t see anything interesting here, isn’t it just a noise? But I’ve added it as a courtesy to Andres Alvarez. Incidentally it’s an example of graph for statistics which I didn’t include in this post ;-)

I may be reading too much out of it but… do players simply try less hard to blocks shots in the 2nd half of the season? By the way, by my method Week 14 is at the end of January.

That’s another sign that maybe we shouldn’t panic too much in the first 4 weeks of any season and complaining about quality of play may be undeserved… in a typical season. This one is different ;-)

Technical note: I calculated FT% based on average makes and attempts not by average percentage from all games… and that’s another noise for me… but I may be wrong ;-)

[I calculated 3P% based on average makes and attempts not by average percentage from all games].
Either I’ve made the same mistake 3 times [possible]… or beginning of January for some reason is a very hot period for three-point shooters ;-)

Finally…

Well, some often repeated statements… are just true.
Players in the NBA do shoot better as season goes along… and that’s a hope for all of us in current season ;-)

Also thanks to this graph we can see that games around Christmas are not your usual matchups…