NBA Summer League: The best of the Spurs’ roster

pb
July 14th, 2010 at 5:23 am

“It is thus unsurprising that championship teams display great defense. There is almost certainly something to “defense wins titles”, but it is more cloudy then your elementary statistics.”

Fair points, overall. Obviously, the cliche, “defense wins titles” was not meant to be a scientific conclusion. That said, the data I presented here does not present a “cloudy” picture as it relates to defense. After all, the top-rated defense was the title winner in FIVE out of the last 12 years, and there has been ZERO top-ranked offenses that have won the title in the past 12 years. That data point alone gives a hint at the superior influence of defense compared to offense in terms of winning titles. It’s only “cloudy” in the sense of lacking a scientific study, but the preliminary data presented herein does suggest the makings of a pretty strong hypothesis.

bduran
July 14th, 2010 at 6:10 pm

“Dif gets screwed because the numbers go 1,1,1,3,4. That’s the problem with median.”

Well, it would help to have a larger data set, like in the next 5-10 years. But it really doesn’t make sense to go back further than 1995, or use teams that aren’t “real” contenders (many semi-finalists & some losing finalists).

“I would like to use Ortg, Drtg differential, but I’m not sure how to get a ranking for this without going to basketball reference and calculating every teams differential and then ranking them for every year.”

Yeah, actually, that’s what I did. It’s a bit time consuming, but that’s how I calculated my “medians” for ORtg. & DRtg. In fact, I reported earlier all the differential rankings for the 12 title teams. Here’s the table again:

E = efficiency differential ; DR = defensive ranking; OR = offensive ranking (*using nba-reference).

98-99 – Spurs – OR – 11th; DR – 1st; E – 1st
99-00 – LA – OR – 5th; DR – 1st; E – 1st
00-01 – LA – OR – 2nd; DR – 21st; E – 8th
01-02 – LA – OR – 2nd; DR – 7th; E – 2nd
02-03 – Spurs – OR – 7th; DR – 3rd; E – 4th
03-04 – Pistons – OR – 18th; DR – 2nd; E – 5th
04-05 – Spurs – OR – 8th; DR – 1st; E – 1st
05-06 – Heat – OR – 7th; DR – 9th; E – 6th
06-07 – Spurs – OR – 5th; DR – 1st; E – 1st
07-08 – Celtics – OR – 10th; DR – 1st; E – 1st
08-09 – LA – OR – 3rd; DR – 6th; E – 5th
09-10 – LA – OR – 11th; DR – 4th; E – 7th

So I can provide the median “differential” for the 12 title winners now, from 1998-1999 up to present, from left to right.

1, 1, 1, 1, 1, 2, 4, 5, 5, 6, 7, 8 = 3.0

For “defense”:

1, 1, 1, 1, 1, 2, 3, 4, 6, 7, 9, 21 = 2.5

For “offense”:

2, 2, 3, 5, 5, 7, 7, 8, 10, 11, 11, 18 = 7.0

When looking at the table above, and all three patterns here, you can see that the median defensive rankings appear on the surface to be in effect pulling the differential rankings down.

Just for example, we’ll delve in a little deeper to the most striking pattern:

“All five of the number one rankings for defense & differential occurred during the same season.”

Okay, so during those years, what was the ORtg. of those title winners?

1998-1999 - 11th
1999-2000 - 5th
2004-2005 - 8th
2006-2007 - 5th
2007-2008 - 10th

Clearly, during those five years, “defensive” ranking carried the “differential” to it’s number one rankings. The rest of the years show a bit more mixed effect on differential in terms of offense and defense, but still, overall, the effect on a differential appears to be “slightly” more for defense.

2000-2001 - greater effect - offense
2002-2003 - greater effect - D
2003-2004 - greater effect - D
2005-2006 - greater effect - O
2008-2009 - greater effect - O
2009-2010 - greater effect - D

Now, I do admit that the Spurs team’s do tend provide a relatively large effect on the data. Their teams during this generation/era appear to give “defense” a consistent edge during this 12 year period. I’m just saying, the Spurs were really THE team, under Pop, that ushered in this new era of defense in the NBA. And I just don’t really see that era ending anytime soon. Although, I do see the dreaded Heat attempting to create a sort of “modern” version of Riley’s show-time Lakers of the 1980’s. However, I think it will ultimately fail, and I know for the Spurs to be the ones to officially thwart that effort, we’re going to have to get back to Spurs-type “D”, while maintaining our top-ten ranking in offense. A tough task, but doable. We MUST however focus on assembling “defenders” at practically every position, or demand from the people we have more effort & focus. I hope Pop’s not getting too soft as he ages!

“At least this data is easy to get and it wouldn’t be too hard look at something like every team who made the conference finals. This would gives us 32 teams instead of 8.”

I know sample size is an issue with this amateur study, but we really can’t look as far back as semi-finalists. Many of them are not really that close championship level (e.g., the 2009-2010, Suns). Finalists “might” be worth looking at, particularly during years where the finals went at least SIX games.

“Maybe you said this before, but why did you chose 12 years? If we’re looking at “modern” teams we have to decide a cut off.”

The main reason is that I wasn’t able to locate “user-friendly” data sets for the information that we needed, or the data didn’t go back far enough (i.e., Hollinger’s). Additionally, the “defensive era” of the modern NBA, in my view, was ushered in during the mid-to-late 1990’s. It probably started with the second three-peat of the Bulls teams (though they had the greatest offensive player on the planet, Jordan, and also Pippen & Co., and then with the addition of Rodman they began to emphasize “D” even more - in their 2nd three-peat more than their first), and then the Spurs really solidified the notion that great defense was critical to winning it all once Duncan came aboard in 1997, Pop’s first FULL season.

bduran
July 15th, 2010 at 5:38 am

“True, but I don’t like to add subjective bias to the sample. If there’s another way to pick the teams that doesn’t have us deciding for each team who was a contender and who wasn’t I’m all for it.”

I understand your issue with subjective bias, but in my view, the only way to limit this problem is to stick to looking at NBA champions (at the most, include finalists). But there’s no way to avoid selecting a cut-off year, whether it’s 1950 0r 1995. And there is a problem with using teams other than the NBA champions in the sample. The NBA champion is often noticeably better than the final’s loser, and particularly the semifinalists. Thus, I think it’s important to keep the focus on the following question: what are those factors that seem to set the champion apart from the rest of the challengers? There’s a lot of really “good” teams, but the champion is often a distinct step above, and we don’t want our results to actually represent the conflation of two similar, but ultimately disparate things; that is, “very good teams that win a lot of games”, and the “NBA champions”.

“…..and maybe everyone who played those teams to six.”

That would clearly dilute our results to the point where we actually end up not measuring the very thing we set out to measure: that is, what are those factors that set apart the VERY best from the rest?

“Looking at the Bulls, the were Ort rankings go 1,1,2,1,1,9. Drtg 7,4,7,1,4,3. So only the last team had better Defense than Offense.”

Yes, but you can see the trend developing toward “improving” their emphasis on defense: an average DRtg. of 2.7 in the three years between 1995 & 1998, versus an average DRtg. of 6.0 between 1990 & 1993. With Michael Jordan on the team, it’s hard for the offense to not be in good shape throughout most of their title years. That guy was simply a scoring machine. But as they added Rodman in 1995, and as Jordan started to move beyond his offensive “peak”, they began to emphasize team defense more, and this simply became more apparent by the 1997 season.

“To me the Bulls were always a fantastic offensive team that played solid D, sort of the opposite of the TD Spurs.”

Opposites within a very circumscribed range perhaps, but certainly not polar opposites. In reality, the Bulls were the “bridge” team between the shift from more offense (1980’s Lakers) to more defense (the Spurs).

“If there is a difference in the game now, I think just looking at the last 10 years will be fine, just need to figure out a way to add more teams.”

I’d be okay with trying 15 years, and adding only finalists that took the champion to at least six games, but that’s about as far as I think makes sense, or we’ll end up measuring something that we did not intend to measure: and that is, what are those factors (defense, offense, differential) that seem to most differentiate the very top teams from the rest of the challengers during this current generation of players?

bduran
July 15th, 2010 at 5:34 pm

Well, the guy made an interesting attempt to use more sophisticated statistics & methodology to try and establish the relative importance certain aspects of the game have on winning games. That said, it’s obviously a work in progress, and of course it says nothing directly about what it takes to win championships, which I suspect often depends on a combination of intangibles in addition to strong box score performance (perhaps some of the 6% variability that he admits is unexplained by box score variables).

“It would be cool if he redid with offensive stats vs defensive stats.”

Yeah, that could be interesting.

“It shows that the defense that the box score doesn’t capture isn’t that important in explaining wins. I suspect this is because good defense is highly correlated with blocks, steals,and def rebounds.”

I agree that it’s generally not that important, but I suspect that some defensive & miscellaneous intangibles may play enough of a role to matter in a tight NBA Final’s series (e.g., “forced” turnovers by the defense [not unforced turnovers or steals], or the tenacity required to get more than your share of “50/50” balls).

Also, the author made this assertion:

“Improving your shooting efficiency is the best thing you can do to win more in the NBA.”

But then admitted in the comment section that he may have jumped the gun here, and would need to revisit this project in the future.

From the comments section on the blog:

* Guy
* July 15th, 2010

* REPLY
* QUOTE

Interesting analysis. However, you’ve mistakenly included three defensive variables as “offense”: DRB, blocks, and steals. If you count those as defense, you will find that offense and defense are of approximately equal importance.

In fact we know this must be true, because the variance of points scored and points allowed at the team level is about the same. Teams vary in offensive and defensive quality by the same amount. So by definition, defense explains as many wins and losses as offense does.

o arturogalletti
o July 15th, 2010
o REPLY
o QUOTE

Guy,

You’re totally right. I’ll probably revisit this at some point in the Future.

Agreed.