When 94% doesn’t mean 94%

I knew it was coming as soon as the buzzer sounded in Game 2 of the Thunder-Nuggets series.

And sure enough, I found it in the first article I read about the game here...

From the Post article:

The numbers say that NBA teams that go up 2-0 in a series win that series about 94 percent of the time.

I like Benjamin Hochman.  I even follow him on Twitter.  But this "stat" has very little to do with the Nuggets’ chances to win this series.  For instance, what on Earth does last year’s Orlando Magic team plowing through an over-matched Charlotte Bobcats team have to do with two relatively even teams playing out the remainder of this series?  Basically nothing.

So what kind of chances do the Nuggets have to prevail?  Here are three different opinions:

First, here are Chris’s and my numbers…

OKC in four - 9.3%

OKC in five - 26.9%

OKC in six - 14.1%

OKC in seven - 23.8%

DEN in six - 12.2%

DEN in seven - 13.7%

So we still give Denver a 25.9% chance to win this series even with losses in the first two games.  Why?  Well, largely because they have three of the next five at home and because we still believe they are a better team (unfortunately, the best team does not always win the series, especially when that team lacks home court advantage). 

 And we aren’t the only ones who think the Nuggets are still the best team in this series.  There is a statistic called Regularized Adjusted Plus-Minus ("RAPM"), which is the most predictive of the plus-minus ventures that we can use to predict the outcome of this series (it uses a four-year sample size because you have to use large sample sizes for plus-minus numbers to mean anything). 

Here’s a quick background on some of the versions of plus minus that are discussed in an NBA context.  The most basic version is the plus-minus statistic itself.  This is just the number of points your team outscores its opponents by during the time the player is on the floor.  So if the team scores 10 points and gives up 8 while a player is on the floor, his plus-minus number is +2, if his team is outscored 13-8 while he is on the floor, he gets a –5, etc…  Adjusted plus-minus is the same thing only a regression is used so that the strength of a players teammates and opponents is taken into account.  Unfortunately, adjusted plus-minus performs poorly in predicting future outcomes, partially because of the need for a large sample, and partially because of the nature of the statistic. For this reason, RAPM was developed using a different form of regression (called a ridge regression).  It performs significantly better from a predictive standpoint than any of the other methods. 

OK, so does this apply to the Nuggets-Thunder series?  The advantage of using this method in the Nuggets’ case is that you can use only the players that have played since the trade, and you can adjust the numbers based on the amount of time each player plays (which we do as well, but not many others do).  RAPM predicts the following point differentials: DEN +5.7 and OKC +4.4.  A team that has a +5.7 differential can expect to compete for a championship (and, not coincidentally, if George Karl would use different lineups, as @chantech suggests, the projection would be even better). 

Using these numbers we can use math to estimate the Nuggets odds of winning this series.  Using RAPM, the Nuggets have a 26.9% chance to win the series.  So, using a completely different method from what Chris and I do, we come to almost an identical conclusion.

There is one more, slightly less valid method we can use to estimate the Nuggets’ chance to win this series.  We can use betting odds to determine an expected winning percentage for the series.  This is a pretty quick-and-dirty method because series prices aren’t as efficient as game-by-game numbers.  But, without boring you about the details, we can use a reputable site’s numbers (DEN +385, OKC -450) to calculate that Denver’s chances are about 20.1%.

So, in these three methods we can safely say that the Nuggets have a much better chance than the 6% number that is typically thrown around.  In fact, I would estimate that the Nuggets have about a 25% chance of coming back to win the series. 

In other words, flip a coin twice; the chance that it comes up heads both times is the chance the Nuggets have to pull off the "miracle."

Write respectfully of your SB Nation community and yourself.

Log In Sign Up

Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Denver Stiffs

You must be a member of Denver Stiffs to participate.

We have our own Community Guidelines at Denver Stiffs. You should read them.

Join Denver Stiffs

You must be a member of Denver Stiffs to participate.

We have our own Community Guidelines at Denver Stiffs. You should read them.




Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.