The Best Ball Big Threes 1986/87 thru 2013/14

Posted on 01/13/2015 by


I like making beautiful and informative things.

A few days ago, I cribbed the idea of a best ball MVP (The Wilt) from Bill Simmons’ weekly NFL column who cribbed it from a reader, who cribbed it from rugby’s Brownlow Medal.  The basic idea is that we give out point to the three players for every single game played (3 points for 1st place, 2 points for second place and 1 for third place) and we then keep track of it and use it to determine the best player in the league. Given that it just so happens that I have a metric that keeps track of every single game (explained in detail here) and assigns a score to each and every player, creating a tool for and keeping track of the Wilt is dead easy. The  Wilt is so named because it’s meant as an objective measure of who can we count on to blow away his peers on a night to night basis and nobody did that like Wilt.

Let’s give you a small update:

Wilt 1 _12

Chris Paul keeps his lead by virtue of the schedule. Curry however has the best balling average at 1.71 (Wilt would be proud).

This is not however what we are actually here for. I happen to have game by game numbers going back to the 1986-87 season and this lead me to do some research. What if I calculate the Best Ball scores for every team from 1986/87 to 2013/14 and do some analysis? What will I see?

For this experiment, I compared multiple variables to the Win Loss record.

R-Square of Variables versus Win Loss %
Percent of Possible Best Ball total points 92.8%
Percent of Possible Best Ball 1st place points 88.7%
Percent of Possible Best Ball 2nd place points 75.9%
Percent of Possible Best Ball 3rd place points 63.1%
Percent of Possible Best Ball points by best player 49.5%
Percent of Possible Best Ball points by best two players 64.2%
Percent of Possible Best Ball points by best three players 73.1%

What’s really interesting is that 93% correlation between percent of possible best ball points and Win Loss percentage virtually matches the correlation between point margin and Win Loss%.

Ignore if you're not a math junkie

Fell free to ignore if you’re not a math junkie

Knowing the percentage of the time a team had the best three players on the court is as effective as knowing their actual point margin at predicting a team’s win loss record. Funky.

If we take it to the three best player’s for each team we see that the performance of those three explains about 73% of what we see in the Won/Loss record. Pareto holds. The Big Two account for 64% of the Win/Loss record on average. The Wilt on each team is at 50%.

Knowing the Big Three for a team and how they relate to the rest of the league tells us most of what we need to know about a team.

Now we get to the good part. With all that handy data, I decided to make a poster showing all the Big Three for every team, by season, from 86/87 to 13/14. The poster shows: the team, the season, the team’s win loss % (WL %), the #1, #2 and #3 players, the percent of Wilt points from the team’s big three, the percent of Wilt points from the rest of the roster and finally the ratio of the too. This ratio is meant to reflect if a team is top heavy or a deep affair.


Big Three Poster

Our top six top heavy teams (as you’ll see), are a mix . It’s the 1989,90,91 and 92 Bulls (2 conference finals and 2 rings) and two young up and coming teams in the 1991 Warriors of Run TMC and the 1994 Shaq/Penny/Anderson Magic.

The roster heavy team of most note? The Defending Champions Spurs in their most recent incarnation.

Enjoy and don’t forget to click to enlarge.


Posted in: Uncategorized