Half-Baked Outliers

Posted on 09/24/2010 by


An outlier is an observation that is different from the rest of a data set. An outlier is an exception to a rule. Typically, in statistics we look for outliers either to protect from their effect where their presence does not indicate false conclusions but rather noise to be discounted or to identify where our basic assumptions and models need revision.

A while back, reader Neal Frazier asked:

I have a question I just thought of about your ‘half baked notion’ – Even though the top 6 produce ~99% of the wins of playoff teams on average, the rest of the team can still be vital to winning. I am imagining a scenario where the playoff pool consists of 2 types of teams, a deep type where 7+guys produce +2 wins collectively and a shallow type that produce where 7+ guys produce -2 wins collectively. The average would be 0 wins produced by the 7+ guys, but the reality for each team individually would be that the 7+ guys would have a big impact on team outcomes.

When you look at the wins produced of the 7+ guys on a team by team basis are they mostly close to 0 or are there teams with strong negative and strong positive production?”

This is a great question about outliers and the short answer is not really, the half baked theory holds for the most part. For the long answer (which follows now) we’ll look a the data . First off, some background.

The Half-Baked Primer

To review , the half baked notion is that what wins in the regular season is not necessarily what  gets you the trophy. Where you need depth (you top 10) to win in the regular season in the playoffs your top 6 is what matters. For the regular season,  using all the data from every season since the merger, the data looks like this :

For the playoffs,  using all the data from every season since the merger, the data looks like this :

So we can conclude:

  • Your starting five account for 82% percent of your wins in the regular season.
  • Your second unit is important over the course of an 82 game regular season accounting for 18% of your wins
  • After that everybody else is statistically meaningless.
  • The best two players accounted for 56% of a teams wins in the Playoffs.
  • The top three players are just below the pareto threshold
  • The next three (4,5,6th man) account for the rest of the positive win contribution about equally.
  • After that everybody else actually hurt teams in the playoffs.

So you need a good 1-10 in the regular season but the better your top six the more successful you are going to be in the playoffs.

But what are there exceptions to this rule? Let’s take a look

Half Baked Outliers:

To look for outliers, I took a deeper look at the full data set for the Playoffs since the merger (all 5877 data points worth of them). I took all players for every playoff team and every year:

  • Ranked/Grouped them based on Minutes played for their team that year
  • Worked out how many Wins they produced for their Team per game (rounded to the nearest .05 of a win per game)
  • Put them in a table
The data looks as follows:
The data approaches a normal distribution (mean is .035 wins per game and standard deviation is .085). If we pick .15 as our point of significance (5 guys at that level mean .75 of a win generated per game and a likelihood of success and it’s also > one standard deviation above the mean), we see that only 14% of the population contributed significantly to playoff success.  Our outliers then are those player who fell one standard deviation above the mean while playing the 7th or 8th most minutes for their team (of which there are 12 out of 1008 players in the data set or 1.2%).  Who are these outliers and did they make a difference?
Looking at the Top30 win generators in the playoffs at the 7th and 8th slot reveals that only four played on teams with a modicum of success in the playoffs (played at least 16 games):
Of the four, Foster playing behind Jermaine O’Neal is the one that swung a series possibly. In general what the analysis confirms is that the theory holds, 99% of the time you’re not going to get any real value from your 7th and 8th man in the playoffs so as a coach you need to play your best guys to be successful .

Right Mike?

Posted in: Uncategorized