Celtics

No NBA Parity and NBA Finals Predictions

| by David Berri

Darren Rovell – on his CNBC.com blog (Sports Biz with Darren Rovell) – offered a discussion today of what the NBA Finals tells us about the lack of parity in the NBA.

For the fifth time in 30 years (27 to be exact), the Boston Celtics will play the Los Angeles Lakers.

It’s the most any two teams in the four major sports have played each other in championship games over the last three decades.

But perhaps more statistically interesting is the fact that over the last 30 years, the NBA has only had eight different champions—the Lakers (9), the Bulls (6), the Celtics (4), the Spurs (4), the Pistons (3), the Rockets (2), the Heat (1) and the 76ers (1).

Popular Video

A police officer saw a young black couple drive by and pulled them over. What he did next left them stunned:

Popular Video

A police officer saw a young black couple drive by and pulled them over. What he did next left them stunned:

Compare that to the other leagues.

Counting this year, the last 30 Stanley Cup Finals, will have yielded 14 different champions.

The last 30 Super Bowls have crowed 15 different teams as Lombardi Trophy winners.

And the league that often gets mentioned the most as somehow having the least parity, Major League Baseball, has produced an amazing 19 different teams as winners over the last three decades.

Rovell then goes on to note…

Take a look at how many different teams, over the last 30 years, made it to their league’s respective final series.

  • NBA: 18
  • NHL: 22
  • NFL: 25
  • MLB: 25

Why is it harder for more basketball teams to make the Finals and ultimately win?

In answering this question Rovell cites The Short Supply of Tall People argument (discussed HERE and HERE in this forum). He also notes, though, that “only 44 percent of teams that have played in it over the last 30 years have managed to win once.”  Percentages are much higher in other sports, suggesting that teams that have won a title before have a significant advantage when reaching the NBA Finals again (relative to teams that have never won a title).

Why would this be the case? Rovell notes that the lack of parity in the NBA may make a seven-game series in basketball more predictable (a point I agree with). He also notes that maybe some organizations are simply better at finding the star players a team needs to win a title. Certainly the success of the Lakers and Celtics suggests that this is possible.

A Final Pick

Regardless of the answer, we are now faced with another NBA Finals contested by the NBA’s most successful franchises (which is good news for fans of the Celtics and Lakers and probably bad news for fans of every other team). And although I have lost the TrueHoop Smackdown, we are still asked to forecast the winner of this series. So I thought I would share my thoughts on the Finals (for whatever these are worth).

The model I employ considers only two issues: The efficiency differential of the two teams and who as home-court advantage. This model led to my victory last year, and has correctly chosen the winning team in eleven of the fourteen series this year. And if we used this model for the NBA Finals, the Lakers would be expected to win in six.

But the Boston Celtics are the Lakers’ opponent. And twice the Celtics have won a series when the model said otherwise.  In fact, the Celtics have defeated the two best teams from the NBA’s regular season in the playoffs.

So now I have a choice. I can stick with the model and pick the Lakers. Or I can remember the words of Stephen Colbert. In discussing George W. Bush, Colbert observed sarcastically that Bush “…believes the same thing Wednesday that he believed on Monday, no matter what happened Tuesday. Events can change; this man’s beliefs never will.”

Such inflexibility is not considered the hallmark of great decision-making. And I suspect, if I pick against the Celtics again, I am essentially ignoring what happened on Tuesday. Consequently, I am going to argue – as I noted was possible early in the playoffs* –that the Celtics we saw in the first half of 2009-10 are the Celtics we are seeing today. And if we assume the Lakers today – with an injured Andrew Bynum – are not much better than the Lakers team we saw in the second half of 2009-10 then the Celtics are to be favored (even if the Lakers have home-court advantage). 

Given all this, I am going to pick Boston to win in six games. 

By the way, the above argument echoes some of what Jeff Ma said at Huffington Post about the balance between intuition and statistical analysis.  And although Jeff is going to win this contest – and I am doing as he has done (picking the evidence that fits the desire to favor the Celtics) – I am not really convinced by the “intutition vs. statistics” argument. After all, how would anyone replicate this approach in the future? Should we just pick the half of a season (or some other data points) that happen to be consistent with our desired pick? This approach just doesn’t seem like something that can be verified with systematic analysis.

That being said, here are two more reasons to pick the Celtics:

1. This is a Jeff Ma-Tribute pick.  He stuck with his team (we have to remember that Jeff justified his pick earlier by noting that he was a fan of the Celtics) and now is the True Hoop Smackdown champion of 2010. So I am picking the Celtics as a tribute to Jeff.

2. I really don’t wish to hear anymore about how Kobe Bryant is the greatest player ever (this might be the real reason for this pick).

History of Using Not-So-Good Models

One last note on this choice:  When I was kid in the 1970s I was a fan of the American League (I grew up in Detroit) but I hated the Yankees (I grew up in Detroit). So I wanted the Dodgers to defeat the Yankees in 1977 and 1978. I then wanted the Orioles to win in 1979 and the Royals to win in 1980. 

In 1981 – after many years of picking wrong (for a kid, four consecutive years counts as “many years”) — I was again picking against the Yankees. But when New York went up 2-0, some kid (I forget who) bet me 50 cents that the Dodgers would come back and win. Of course, I knew better. The Yankees – as I “knew” – were not going to lose to the Dodgers. So I took this bet and thought I was going to make some easy money. Right after I made this bet, though, the Dodgers won four consecutive games and I was out 50 cents.

All of these picks were made without the benefit of a statistical model. Well, actually I had a model.  Bet on the American League and against the Yankees (unless the Yankees are up 2-0 in the series). So it might be better to say that I made these World Series picks with a not-so-good model. And when I look back at my justification for my Celtics pick, I think that once again I am using a not-so-good model. Therefore, fans of the Celtics – given my history with not-so-good models – should be a bit worried.  In sum, I think I am about to lose my “50 cents” again (and once again, have to hear about the greatness of Kobe).

- DJ

*-Here is what I said in discussing the Boston-Cleveland series: “The question,” says Berri, “is if Boston can replicate what we saw in the first half of the season. If that happens, this series can go to seven [and Boston might win]. If that isn’t true, Cleveland takes it in four or five.”