Tis the NBA playoff season, or the time of year when NBA observers enthusiastically draw sweeping conclusions from very tiny samples.
Here are a few thoughts to keep in mind as we all watch playoff basketball over the next two months:
1. As I noted last May, NBA players are essentially working for free in the playoffs. Players have already received their salaries for the year. The monetary reward from the playoffs comes from a) the NBA’s playoff pool (which is quite small relative to the salary many players receive) and b) the potential for a better contract down the road (and I am not sure the playoffs is that important for most free agents). So when you see players give their all in April, May, and June, remember, they are not really being paid to do this. These players are really trying hard because they really like playing basketball (and of course, winning).
2. Playoffs are for fun, not for science. Last year the Golden State Warriors defeated the Dallas Mavericks. It was quite clear from the 2006-07 season that the Mavericks were the better team. But in one week a player(s) and team can play much better (or worse) than expected. This doesn’t “prove” that the winning playoff team is better. In sum, a small sample is not going to trump a large sample.
3. Despite point #2, announcers and sportswriters are going to declare each winning team to be the greatest. And each losing team is going to have obvious holes that need to be filled. The role of luck and chance are, as a rule, generally ignored.
The Best Celtics Team Ever
The playoffs are believed to identify the best NBA team. But as noted, the regular season is a better sample than the playoffs. And when we look at the regular season we already know the identity of the best team in the NBA. And that team is the Boston Celtics.
The Celtics had a truly historic season in 2007-08. How historic? Last June I wrote a column examining the Best Team for Each Team.
In this column was the following table:
Table One: The Best Team for Each Team
We can measure efficiency differential (offensive efficiency minus defensive efficiency) all the way back to 1973-74 (which I have done). With data in hand we can determine the best team in the history of each franchise (at least back to 1973-74). For example, the best Celtics team – prior to this season — was the 1985-86 edition that posted a differential of 8.94. This year the Celtics had a differential of 10.95. This mark easily tops anything we have seen by Boston since 73-74. And given the scoring differential of the teams prior to this date, I think the efficiency differential we saw this year was the best mark in franchise history. Yes, the Celtics with Kevin Garnett, Paul Pierce, Ray Allen (and Rajon Rondo, Kendrick Perkins, James Posey, Leon Powe, etc…) is the best team in this historic franchise’s history.
Of course some will say, “when this team wins a title (or three), then they can rank among the greats in Boston history.” And to those people I say, “No, no, no. Did you not read the first part of this column? The sample from the playoffs does not trump the sample from the regular season. And the regular season sample we have from Boston trumps every other regular season sample we see in this team’s history.”
And I will add… even if Boston falters in the playoffs, this team is still the best in franchise history. Again, the regular season sample is better than the playoff sample.
Other Best Teams
By the way, the Celtics are not the only franchise to post their best mark in team history in 2007-08. The following teams also offered their best this year:
Detroit Pistons (8.17 differential trumps the 7.40 mark seen in 2005-06)
New Orleans Hornets (5.70 differential trumps the 3.47 mark seen in 1994-95)
Toronto Raptors (3.11 differential trumps the 2.37 mark seen in 2000-01)
Two other teams came close to setting a franchise best record. The Denver Nuggets had a mark of 3.70 in 1987-88. This year their differential was 3.59. And Houston had their best differential last year. This year’s differential of 5.02 fell just a bit short of the 5.19 mark we saw in 2006-07.
And for those who are interested, here is every team’s efficiency differential this year.
Table Two: 2007-08 Efficiency Differentials
Tomorrow I am going to tell another story from the regular season sample that just concluded. This story is going to focus on the NBA’s worst.
A WoW Contest (sort of)
Let me close by announcing “sort of” a WoW Contest (and I will explain the “sort of” part in a moment). I would like to know which announcer and/or sportswriter reaches the most dramatic conclusion in the playoffs from the smallest sample.
In the comments, please post your observation. The “best” observation will win… okay, nothing. I don’t have anything to give away. That’s why this is only “sort of” a contest. I will try and note the winner in a forthcoming column (yes, that’s not much of a prize, but it’s all I got).
– DJ
Our research on the NBA was summarized HERE.
The Technical Notes at wagesofwins.com provides substantially more information on the published research behind Wins Produced and Win Score
Wins Produced, Win Score, and PAWSmin are also discussed in the following posts:
Simple Models of Player Performance
What Wins Produced Says and What It Does Not Say
Introducing PAWSmin — and a Defense of Box Score Statistics
Finally, A Guide to Evaluating Models contains useful hints on how to interpret and evaluate statistical models.
Wages of Wins Journal 
yow
April 21, 2008
“And to those people I say, “No, no, no. Did you not read the first part of this column? The sample from the playoffs does not trump the sample from the regular season.”
Unfortunately, I’m pretty sure that very few people will share your sentiment – especially those who play/coach the game. Honestly, who cares about regular season records? Have you ever seen an actual NBA player who’s proud because of his team’s regular season ‘efficiency differential’? And whoever decided that efficiency differential should take precedence over everything else (including the actual regular season record and playoff runs)?
PJ
April 21, 2008
Could the weakness of the Eastern Conference– apart from Detroit and Boston– have inflated somewhat the efficiency differentials for Detroit and Boston? Both teams are excellent this year, but it seems unlikely to me that this year’s Celtics would be notably better than the 1985-86 team, or that this year’s Pistons would be the best Detroit team ever. (How do the late 1980s Pistons match up?)
stephanie
April 21, 2008
“2. Playoffs are for fun, not for science. Last year the Golden State Warriors defeated the Dallas Mavericks. It was quite clear from the 2006-07 season that the Mavericks were the better team. But in one week a player(s) and team can play much better (or worse) than expected. This doesn’t “prove” that the winning playoff team is better. In sum, a small sample is not going to trump a large sample.”
Last year Golden State swept the season series and systematically dismantled the Mavericks in the post season without the benefit of home court advantage. How many games would they have to play against each other before we can say that yes, Don Nelson is a better coach than Avery Johnson in this situation and 06-07 Golden State presents serious match up problems with the Mavs? Maybe the Mavs are better against other teams than GA (hence the regular season records) but GA has the Mav’s number for whatever reason. Isn’t that possible?
On paper, the 06-08 Mavs are an all time great team. I really like the whole stats things and I’ll defend it against anyone who says certain conclusions don’t feel right. But excuse me if I have a difficult time swallowing the 06-07 Mavs as a team for the ages.
Tim
April 22, 2008
How about this:
Finals Countdown: LeBron soars above adversity, McGrady wilts
Steve Greenberg, SportingNews.com
http://www.sportingnews.com/yourturn/viewtopic.php?t=403039
Key quotes:
“As Michael Jordan did before the Bulls had a decent enough supporting cast to become true contenders, James makes the Cavs dangerous by himself. ”
“Maybe it’s harsh to blame McGrady outright for his failure to advance past the first round in his career, but at some point the guy has to rise to the occasion. All the great ones do. McGrady’s team is overmatched, no doubt, but his routine of morphing into his cousin Vince Carter at money time is unacceptable.”
Is that the sort of thing you had in mind?
Frederic
April 22, 2008
Hey PJ, I think a good answer to your question is in a recent post from last week of this blog: “Does it matter who you play”. Of course in that post only winning and loosing records are compared, but I would be surprised if it is a totally different story with efficiency differential.
Brent
April 22, 2008
This will come out of someone’s mouth:
“With the Celtics loss in 7, all of the work and energy put into this season and all of the machinations of general management to obtain Kevin Garnett and Ray Allen were for naught.”
For that matter, I already here people saying, “Doc Rivers is a shoo-in for coach of the year” – Yep, he sure did a much better job coaching this year. I can’t think of one other thing about the Celtics that could have changed from this year to last year and driven the dramatic improvement in wins and losses. Certainly, his coaching was the difference! Doc is brilliant!
Tball
April 22, 2008
DJ,
I think the Efficiency Differentials chart is mislabeled. It should read points scored per 100 possessions. Not that anyone will be confused that teams are scoring 100 points per possession, but clarifying the chart will allow more people to appreciate it.
Tball
Jason
April 22, 2008
It doesn’t appear to me that an unbalanced schedule playing more contests in the east was responsible for Boston or Detroit’s efficiency differential. Both teams did very well in wins and losses against the west, about as well (or better) than we’d expect given their overall record and both teams have efficiency differentials that appear in line with their overall record compared to the rest of the league.
Owen
April 22, 2008
Not quite what you are looking for, will keep my eyes peeled, but I love this quote from the Rocky Mountain News’ Dave Kreiger….
“This could turn out to be an epitaph for the current edition of the Nuggets. Lots of good things. Iverson and Anthony each put up 30 points and Linas Kleiza and J.R. Smith combined for 38 off the bench. That should be good enough to win.”
PJ
April 22, 2008
Discussing this article at another website, and another commenter made what seems to me a shrewd point:
“The more accurate statement for [Berri] to make would be ‘Best Celtics regular season of all time with respect to the rest of the league.'”
Isn’t that actually what you’re measuring here (not Best Celtics Team of All-Time)?
Given that this was, according to efficiency differential, the best season by the Celtics in any year relative to the rest of the league, then this team almost certainly deserves to be in a conversation about best Celtics teams of all-time. But it seems to me that to actually have that conversation, you would need to introduce other factors (strength of schedule and strength of league, firstly and secondly, and then perhaps performance against the other elite teams in the league that year, etc).
JChan
April 22, 2008
Yow has a good point. While efficiency differential is probably the best indicator of how good a certain team is….that will be small consolation to the folks in Boston if the Celtics don’t win it all.
Fans want championships. And if the Utah Jazz manage to go all the way this year, I won’t care that they were only fourth in efficiency differential, I’ll proudly raise my finger and chant “We’re number one! We’re number one!”
Of course, if they lose to the Lakers, I may take some small consolation knowing that they lost to a better team.
Taylor
April 22, 2008
Nothing, hyperbole-wise, can trump Jim Rome’s declaration on his radio show last spring that Baron Davis should have been the NBA’s 2006-2007 MVP. I would like to extend Mr. Rome’s reasoning to its logical conclusion, which suggests that Kyle Korver went from being a horrible “clutch” player to a great one in the split second that it took Andrei Kirilenko to rebound Korver’s shot from the side of the board. “Reasoning,” indeed.
Alex
April 22, 2008
Pretty cool for Pat Riley that he coached three of the “best of” teams (NYK, MIA, LAL)
yow
April 22, 2008
Imagine that the rules were changed such that the team with the best efficiency differential is declared the best team. How the games will be played very very differently..
Wins and losses won’t really matter – teams that are losing will fight to the end, whether they’re losing by 1 or losing by 10 because they want to minimize points differentials. Basically, there will be no insignificant games. Winning by 40 points against an inferior opponent would be much more important than beating a quality team by 1 point.
Animal
April 22, 2008
Dave, you can’t just look at stats. You need to factor in the human element. Humans play basketball not numbers.
Christopher
April 23, 2008
Is there a website where you can get efficiency stats that are sortable with splits etc (all at the same time)? I can’t find one.
Ryan
April 23, 2008
Could somebody explain to me how possessions are kept track of for the efficiency statistics. Can you derive this information from stats found in the box score?
Christopher
April 23, 2008
Followup: Can you look at the modern era of NBA basketball and see how well the “best” team, in terms of efficiency diff., fared? How often do the “best” teams win the tourney?
Stu
April 23, 2008
My discaimer – I have been a die-hard Celtics fan since the mid seventies.
I am thrilled with this year’s team, but this is a perfect case where the stats don’t tell the whole story. If you had to bet your life’s savings on the outcome of a playoff series between the ’85-’86 Celtics and the ’07-’08 Celtics, which team would you bet on? I wouldn’t think twice about putting all on the old guys. Anyone who says they’d put their money on this years team is lying. If that doesn’t make you realize that the stats are misleading, try going down the rosters one position at a time and tell me who you’d pick at each spot. There are a couple of spot where this year’s team would get the nod, but just barely. For instance, I would take Garnet over McHale at PF, but just barely. But at SF, Bird beats out Peirce with his shooting hand tied behind his back (sorry Paul, I still love you). Again, Ray Allen is better than Danny Ainge, but the difference between Robert Parish and Kendrick Perkins is so subtantial that the ’85-’86 team would win in a landslide. And if you are still not convinced, let me remind you that this year’s team doesn’t have Bill Walton coming off the bench.
Evan
April 23, 2008
Stu —
Not that it matters, since we can’t prove it either way, but I’d bet on this year’s team.
Generally, when offered a bet with positive expected value, I tend to take it. But then, I like money.
Owen
April 23, 2008
Stu, what about Rondo, Posey, and Glen Davis?
I think there is a tendency to exaggerate the greatness of the past. I would put my money on this year’s team I think also…
Tim Ricchuiti
April 23, 2008
You say that the regular season is a “better sample” than the playoffs. The question is, better sample for what? I would argue that we need to make sure the sample and our conclusions match up. The regular season is a better sample than playoff basketball in terms of making conclusions about regular season basketball. By any measure, the intensity, the quality of the teams, and the pressures, risks, and rewards involved in playoff basketball are significantly higher than we find during the regular season. Therefore, comparing efficiency differential, or any other metric, between the regular season and playoffs is a relatively useless endeavor (unless your goal is simply to quantify the difference between the regular season and the playoffs).
Beyond which, there’s quite a difference between saying something like “the playoffs are believed to identify the best NBA team” and something like “the playoffs are defined and designed to identify the best NBA team.” If the NBA announced that it would cease having playoffs, and determine its yearly championship instead based on, for example, efficiency differential, than basketball organizations would design their teams significantly differently, and coaches would coach the regular season and players would play the regular season significantly differently. It’s because the playoffs are the method by which the NBA chooses its champion that playoff basketball is that much more important (and treated more importantly) than regular season basketball.
Costa
April 23, 2008
The way I see it, there’s no reason to segregate the two. One can make an opinion on the quality of a team by looking at both the regular season and playoff records. Sure, that means that we’re not comparing the same number of games, but how bad is that? This is not an either/or situation.
Why not look at the team’s efficiency diff combining both the regular season and the playoffs? Dismissing the sample of games played in the playoffs, which can be as big as 25 games, seems almost as short-sighted as dismissing the regular season. The playoffs are making each team (that makes it)’s sample bigger, and thus more representative.
Integrate the 85-86 Celtics playoff games into their differential, and do the same for this year’s version once the playoffs are done, and we can make another comparison at that point.
Ideally, figuring strength of opposition would be good too.
Jason
April 24, 2008
Ideally, one would look at the standard error of the efficiency estimates as an indicator of how reliable a comparison between the two would be.
As for whether or not the general consensus is that that the 80s edition was better, that’s also a matter of psychology. We’ve been well conditioned to regard that Celtics squads as “among the very best teams ever”. That’s both true, but we shouldn’t discount the effect of history and legend making. In the moment, it’s tough to elevate any team to that status and nostalgia tends not to be the best decider. Popular opinion can be wrong.
In the 80s we’d known about the Celtics from how they’d performed over several years. They’d been very good already, having already won a championship with the same core of players. This also elevated perception. We’re still in mostly uncharted territory with this year’s edition. The preseason over-under for Boston was a win total in the 50s. This suggests that people thought they’d be good, but few expected them to be *this* good. That also tends to produce a bit of a ‘hangover’ effect where I expect some still seem to think of them as something of a fluke to be this good so soon after being so bad a year ago.
I’m also not a fan of the “player at x position is better than opponent at x position” as a means of evaluating a team. That works in matchplay golf, but not so much in basketball.
Bob
April 30, 2008
This year’s Celtics may be the “best in history” in terms of statistics versus their opponents in the regular season, but that’s it. Calling a team the “best in history” without these qualifications is ignoring the fact that playoff games – the games that really matter – are played under very different conditions than regular season games.
Back when there were best of 5 series, I think there were some flukey wins, but I think most everyone – players, coaches and fans (everyone except statisticians maybe) – agree that, assuming everyone is healthy, the winner of a 7 game series is truly “better” than the losing team regardless of how they performed in the regular season.
Crediting “luck and chance” ignores fact that in the playoffs are different. Everyone is playing at an intensity which is not consistently sustained while accumulating stats throughout the regular season. Some players crank it up a notch and thrive with this added pressure, others do not.
It ignores the fact that the officiating is also very different. Much more physical play is allowed which favors stronger, more physical teams.
Lastly, it ignores the fact that you’re playing the same team up to 7 games in a row means that there can be a real mental chess match between the coaches with adjustments to plays & lineups that just doesn’t exist in the regular season.
VS
May 12, 2008
“assuming everyone is healthy, the winner of a 7 game series is truly “better” than the losing team regardless of how they performed in the regular season.”
LOL. IT’S THIS!
THIS is the most ridiculous statement of the post-season.
Thanks for winning me that award man!