Coaching & Analytics: Newsletter #1
From the archives: V1 of the old Hoop Vision coaching newsletter
NOTE: This post was originally sent as an email to subscribers in August 2016; the Hoop Vision newsletter and operation was far different in that period…enjoy:
Welcome to (unofficially) the first-ever coaching analytics newsletter.
You don't have to look very hard to find coaching newsletters with X's and O's, motivation, and leadership advice; in that regard, an analytics-focused seemed only right. Some quick info about me: I am Video Coordinator for New Mexico State University men's basketball. I also have run a college basketball analytics blog, Hoop Vision, on and off since my junior year of high school.
The newsletter was conceived with Division 1 coaches and programs in mind. However, I was surprised and thrilled to get a very diverse group of folks in basketball who expressed interest — from fans to coaches to NBA executives. Let’s go.
Stats Don't Lie?
In some basketball circles, stats and film are viewed in conflict with each other. Getting away from this mindset is probably the first step in better understanding how to utilize analytics as a coach.
In a very superficial sense, neither stats nor film ever “lie.” Rather, they simply describe what happens.
If a player scores 40 points in a game, he scored 40 points in a game. We have the box score to prove it. We also have the film to prove it, and that's simple enough to accept. However, when we start using advanced statistics it can be more difficult for a coach to understand exactly what’s being described. Adjusted defensive efficiency, for example, is admittedly a little more difficult to understand than points. But this is no different than someone with no scouting experience trying to watch film and understand if a player is properly stunting when one pass away at the nail.
Things actually start to get tricky when we use stats and film to predict.
The common rebuttal to stats/analytics is that you can find a number to tell you anything, and often times various statistics can outwardly contradict each other. Yet, isn't the exact same thing true for film? I can find 10 defensive clips that makes Player X look like the best pick-and-roll defender in the world. I can also find 10 defensive clips making that same player look like the worst pick-and-roll defender in the world.
Problems don't stem from the stats and the film, problems stem from the people using the stats and the film.
In both cases, you have to be well trained and extremely cautious when moving from a description (Player X scored 40 points) to a prediction (We need to shrink the floor on all Player X touches in order for our defense to be successful).
A great scout can take a given amount of film and get every last piece of useful information out of it. Along the way discarding the unimportant sequences. A great analyst can take a given amount of data and get every last piece of useful information out of it. Along the way discarding the unimportant numbers. The "new-age" basketball thinker can ideally use both types of information in harmony with one another to even further help the decision making process.
Knowing the limitations of a statistic is extremely important, but there are also some inherent advantages to data-driven decision making. Stats are "watching" every player for every team during every game. There's not enough time in the day to watch every college basketball game. With the the large number of teams and unique styles of play in NCAA basketball, analytics can help keep track of the 351 team and 4,563 players. This is also true for recruiting. With the rise of AAU sneaker circuits and transfers (discussed in more detail below), stats can help with evaluations when your staff only gets limited time to see a recruit.
Transfer Analytics
Sports Illustrated writer Luke Winn has written about the rise of the "Up-Transfer," the term for a player from a low-major league transferring to a mid-major or a mid-major player transferring to a high-major.
As the sample size for these transfers increases each year, we can use historical stats to learn more about these potential transfers. In other words, how do players from Conference A perform after transferring to Conference B? Every player is different, of course, but gathering all possible information is crucial to making recruiting decisions.
The graduate transfer may be the scenario most answerable via analytics. Grad transfers play right away, eliminating the uncertainty of projecting how much a transfer will improve in his redshirt year. They also usually have three years of D1 statistics under their belt.
This offseason in particular featured a few grad transfers that became highly coveted targets despite coming from relatively lowly conferences. To give a very brief (and admittedly, incomplete) example of how data can be a tool for recruiting decisions in your program, take a look at the table below of past MEAC grad transfers:
The sample size here is tiny, and there are certainly better ways of evaluating players than the selected statistics. However, you can see that MEAC grad transfers have not performed well in the past.
Truthfully, this example was left incomplete on purpose. We can only give so much information away without losing the competitive advantage that data gives us at New Mexico State. That's the real takeaway: Anywhere there is data, there is an opportunity to get a leg up on the competition - scouting, recruiting, player development, and much more.