The Perception of the Third Quarter

UPDATE:  I have added the data for this year up to game 17 for the Spurs.  Like CapHill said in the comment section the Spurs are having serious issues in the 1st and 4th quarters if we are looking at the point differentials. 

In an effort to learn how to collect data from the internet (screen scrape, compassionately of course) I decided to take on a small (initially) project and write a fanpost for PTR.  The motivation for this post originated from comments like the following:

THIRD QUARTER SUCKED!!!!! -- Artis Gilmore

I HATE THIRD QUARTERS -- Sleep Research Facility

Ugh. And the Spurs are showing why they have the third quarter collapse under copyright. -- Tim C.

Third quarter collapses are "in" again. -- LatinD

What a depressing third quarter. It’s . . . it’s like watching a Spurs game! -- Lauri

And there are many more colorful comments regarding the infamous third quarter.  So I ventured to find out if the Spurs' third quarter calamities are the real deal or if they are imaginary psychological junk.  A few disclaimers and notes before we begin.  This is the first time I do something like this and although I have randomly checked the data to verify the accuracy of it and make sure there were no mistakes, I did not check the data like I would if I were to publish it in a scientific publication so it comes with no guarantee of accuracy.  Also, these are the most basic statistics possible and there are no adjustments for pace or any other advanced metric.  In addition, I ignored for the time being performance during the playoffs, overtimes and the difference between normal games and SEGABABAs which would add some complications to the data collection.  Finally, if anybody is interested in getting their hands on the entire data set (it is BIG!) let me know and we can work something out.

My initial plan involved looking at the Spurs numbers during the last 10 years to see how their performance in each quarter varied per year.  If you are keeping count that would be 10 years x 82 games x 8 scores (4 quarters for Spurs and opponent) which yields 6560 data points collected.  Of course when I was done doing that I realized that the results dont mean anything if you dont compare them to the rest of the league, this led me to collect 6560 points x 30 teams for a total of196800 points! whew! Im glad computers are good with numbers.

OK, enough rambling and lets look at the data.  The following plots show the league minimum, maximum and average points scored per quarter along with the same number for our Spurs for every year in the data set.

That was fun wasnt it?  One of the main features is that the Spurs have been right near the league average offensively in every quarter.  What is more incredible is that the third quarter has been notoriously close to the league average and it does not show a lot of variation, something we would not expect from our understanding of the third quarter performance of our team.  A few additional noteworthy observations include the difficulty of the team in scoring point in the 4th quarter the last three years indicating a possible difficulty in closing out games, and the scored points per 1st quarter in 2007-2008 which was near the league minimum, which might indicate that the guys were behind after the first in many of their games (more on that later).

Now that we looked at the offense lets take a peak at the defense.  And the defense is a thing of beauty for most of the decade.

WHOA!!!  That is some nice defense.  The first thing that jumps at me is our defensive domination compared to the league in the first and second periods.  Those are some beautiful data points.  In the third quarter we were still near the league minimum but somewhat closer to the league average.  The fourth quarter went up and down throughout the decade.  A second interesting observation is that we had the best points allowed per quarter the last three years in the fourth quarter which for me was unexpected.  There is not much more to say about the defense by itself just like the information in the offense is somewhat ambiguous.  A better approach is to look at these together and calculate the average points differential per quarter.  Here a positive number is good and a negative number is bad.

OK.  Guess what is the league average point differential per quarter?  That is right, Im a genius and spent a bunch of time calculating a quantity that is obviously zero.  In any case the exercise had didactic consequences and I dont regret doing it.  In any case this is one of the more revealing findings.  Look at the first and second quarter data and see how we used to essentially get solid leads for the early part of the decade.  In 2007-2008 and 2008-2009 however the point differential for these quarters decreased dramatically.  In some instances there were probably as many of those quarters in which we had a lead in those years as those were we didnt.  Not a good omen for dominating basketball games.  In 2007-2008 especially we were behind after the first quarter as much as we were ahead (probably) which means that the team had to fight back in half of their games.  A nice year was seemingly 2002-2003 in which it appears that the team was so dominant in the first three quarters that they could ease it in the fourth (and maybe rest the starters? or maybe almost blow it in the fourth? some Spurs veteran fan can tell us).

To finalize, the main objective of all this was to find out if the third quarter is really worse than the other ones for us.  I think the answer is partially answered in that it does not seem that the third quarter has ever been any worse than the others.  If I had to call any quarter as the Spurs worst would be the fourth in which our defense and points differentials are all over the place.  As for the reasons we have this perception, maybe further analysis would reveal the answer.  A possibility that I have thought about is that maybe some of the third quarter collapses have been huge.  Its a possibility that can be verified with the data set I have built and I hope I can do it (or maybe some of our resident statisticians) before the seasons is over.

Thank you all for reading! If there is enough interest I can make a monthly post monitoring the numbers for this year (I already have a little script that would do most of the work) let me know if you would enjoy that.