There isn’t a fan alive who hasn’t thought or uttered these words
“If I were a general manager, I would do [this hair-brained trade idea]”.
With programs like the NBA Trade Machine, you can now go to ESPN and other sites across the web and simulate hypothetical transactions. Some deals work and some deals don’t. But in the back of your mind, you believe that you have the answer for all that ails your team.
If they would only listen.
After spending time earlier this month in Boston for the annual MIT Sloan Sports Analytics Conference, I came to this conclusion: if ever handed keys to the kingdom, my first order order of business would be hiring someone much smarter than myself.
I can read body language, observe skill sets, and give honest reads on player personalities.
But in today’s NBA, that’s not enough. You need more.
You need a young man like Robert Ayer, who currently works for a high-tech start-up and not in professional sports. The lack of first-hand hoops experience didn’t stop Mr. Ayer from spending time and energy on research titled “Big 2’s and Big 3’s: Analyzing How a Team’s Best Players Complement Each Other”.
First, let me get some of the easy stuff out of the way. “Big 2’s and Big 3’s” is the term that Ayer attributes to the combination of a team’s primary players. The Miami Heat’s LeBron James, Dywane Wade and Chris Bosh are an example, but every NBA squad has a big two or a big three. They just vary in skill level.
Let’s roll up our sleeves and dig in.
Several issues should be addressed before running this regression. The first is how to quantify team talent. Most observers have a good idea of how talented a team is, but quantifying this variable is less straightforward. The approach taken was as follows. First, each player-season was evaluated in terms of the player’s efficiency that year, measured by the efficiency formula used on NBA.com: ((Points +Rebounds + Assists + Steals + Blocks) – ((Field Goals Att. – Field Goals Made) + (Free Throws Att. – Free Throws Made) + Turnovers)). Players were then classified according to their season efficiency rating, based on the percentile rank of the player’s efficiency rating (compared against player-seasons across all seasons), and points given to each player according to the scoring system below, weighted by minutes played. For each team-season, their player’s minutes-weighted score is summed, giving each team a metric that estimates the talent of the team.
Got it? These player efficiency ratings were then fed into one of two formulas.
Model 1: Wins = Team talent + Coaching + Composition of top 2 players
Model 2: Wins = Team talent + Coaching + Composition of top 3 players
Notice that Ayer included a variable that measures coaching as well. For our purposes, we are going to gloss over the coaching numbers, but rest assured, the research is there.
And here is one other piece of business that assuredly someone might ask: Ayer inserted the numbers of every player from 1978 to present. There is no small sample-size warning to adhere to.
Why is this information relevant if you’re a Sacramento Kings fan?
Pretty simple. The Kings have a big three (Tyreke Evans, Marcus Thornton and DeMarcus Cousins), but they seem to be a wrong fit. We can try to make an educated guess why, but it would be just that – a guess.
Ayer has made a road map. If the fit is wrong, why not look for the right match? Why not assess which one or two of the Kings’ “big three” work together historically, then use the other leg of the triumvirate to acquire a player who statistically fits with the other two.
If the data is out there, isn’t it worth exploring?
According to Ayer, there are 13-different player types. The player types are pretty much what the common basketball junkie would come up with, but here is an example.
Cluster 14: Role-playing big men without an exceptional skill, but contribute in several categories: Udonis Haslem, Kurt Thomas
(Jason Thompson, anyone?)
We need to squint on a few of these in regards to the Kings’ big three, but here are my best guesses.
Cluster-2: High scoring, dynamic guards (mostly 2 guards, but some 3’s like Grant Hill), typically not great 3 point shooters, or if they are, don’t shoot very many: Kobe Bryant, Dwayne Wade, Tracy McGrady, Adrian Dantley.
Cluster-4: Wing 3-point shooters: Dan Majerle, Shane Battier
Cluster 5: Dynamic, well-rounded power forwards, strong rebounding, dynamic 3’s: Chris Webber, Pau Gasol, Kevin McHale
It is possible that Cousins could become a “Cluster 12” player (Ayers cites Hakeem Olajuwon and David Robinson as examples, but Cousins isn’t in the same neighborhood defensively as them). Thornton isn’t a perfect “Cluster 4” either, but he’s close enough.
The data shows us that if Cousins is the Kings’ center piece, which I believe we can agree on, then there are two specific “big three” cluster groupings that work particularly well. The Kings should be looking for a 2-2-5 pairing or a 2-5-8 pairing.
We know what a 2-2-5 pairing is and the Kings have two of the three pieces in place (Evans and Cousins). In this model, they would need to look for another Evans-type player. Think James, Wade and Bosh, if it makes things easier.
A 2-5-8 grouping is intriguing as well. Here is the description of a “cluster 8” player type:
Cluster-8: Multi-faceted, high scoring wings, with high assists for their position and are great 3 point shooters: Paul Pierce, Danny Ainge
This is where it gets tricky. Where can we find these player-types and do the Kings have the pieces to land one? “Cluster 8” just so happens to be a difficult find and are extremely valuable. What they should be looking for is a player like Marcus Thornton, but with a much higher assist rate.
A lot of the players that fit this mold are at the tail end of their careers and are possibly unattainable (Pierce, Bryant, Ray Allen). You can add a Joe Johnson to this list, but neither the Kings nor any other team is taking on that contract.
Either way, we have a blueprint that might work. We have data and mathematical equations that tell us something. They might even give us the formula for finding the perfect fit to put alongside DeMarcus Cousins and possibly even Tyreke Evans.
So if I were handed the keys to the kingdom, I would hire Robert Ayer – a smart kid. I would hire him and between the two of us, we might look for a few more smart kids, because there are things we know, but a whole lot that we don’t know either.