August 24, 2012 in Cover Story, Hoop Convos
There are few things better than getting philosophical about the game of basketball. In ‘Hoop Convos,’BallinMichigan will do just that, engaging in (hopefully) meaningful, free-flowing conversations with compelling writers and thinkers who love the game as much as we do. This week, we’ll have three special conversations, all related to the advanced stats movement in basketball and its ties to the state of Michigan.
In our third and final ‘Hoop Convos’ post related to the state of Michigan’s connections to advanced statistical analysis in basketball this week, I’m featuring a conversation with Ben Gulker (), a writer who has helped fill a niche by covering the Detroit Pistons for Detroit Bad Boys, SB Nation’s Pistons news site, using advanced stats for several years.
But the reason I asked Ben to participate in this series goes deeper than just the fact that I admire his writing. He’s also a former college basketball player, as he’ll describe below, so I wanted to get his take on the perception that advanced stats are just for the numbers nerds with no athletic ability.
One of the things I hope these conversations this week help illuminate is that advanced stats are not scary or threatening or nerdy. They’re simply another way to continuously look at and evaluate the game we all love to come up with the best possible methods for evaluating and explaining what actually happens on a basketball court. Ben has a lot of interesting insight worth reading below and you’ll also get to find out which NBA All-Star he once hit a floater over.
First, when and where did you play college ball? Any accomplishments you’re particularly proud of, including scoring on Chris Kaman?
I played college ball for three years at Spring Arbor University (SAU) as a freshman, sophomore and junior (01-02, 02-03, 03-04). I didn’t play my senior year for a variety of reasons, the two most important being several ankle injuries and a growing interest in academics. SAU is a small school and part of the , as opposed to the NCAA. A major difference between NAIA Division 2 and NCAA Division III is that NAIA Division 2 schools can award athletic scholarships, but teams from those divisions often play each other. For example, we competed against Hope College, Kalamazoo College, Calvin College, and similar schools in the state and region.
I am most proud of being voted co-captain as a junior, along with one of our seniors. That my teammates believed in me as a person, teammate, and friend is more important than anything I ever accomplished on the court as a collegiate player.
I am also very proud of my high school team and its success, without which I may not have had the opportunity to play in college. In two years together, our team went 42-5, including going a perfect 20-0 my senior season.
If I had to point to individual accomplishments, there are a few that stick out. My best overall college game came my sophomore season, scoring 18 points in 15 minutes on perfect shooting. I also put up 9 points in about a minute and a half against Albion College as a junior. We were down by 18 when I entered the game and only 9 when I exited, giving us a chance to win. I also got to play against Chris Kaman’s teams in both high school and college. I scored over him on a running floater in a summer league in high school, and played against his Central Michigan squad that made their deep run in the NCAA tournament. That I can say I put up 7 points against that team still brings a smile, because we were totally and completely overmatched by that team.
When did you first start getting into advanced stats? Anything in particular that drew you in?
I always paid close attention to statistics, but not necessarily of the advanced variety. As a kid, I paid lots of attention to my stats, which I think is pretty common at that age. My high school coach, though, helped broaden my perspective. He was particularly interested in how our team collectively and players individually drew charges, deflected passes, generated steals, avoided turnovers, collected offensive rebounds, and shot efficiently. Those weren’t necessarily the things I was accustomed to focusing on. We didn’t have an all-in-one metric or anything, but we were taught – and subsequently learned from experience – that if we did those things well as individuals, we’d get more playing time, and if we did them well as a team, we’d win games.
It wasn’t until sometime in 2008 that I got into advanced stats, when I was introduced to Dave Berri’s book, The Wages of Wins.
Pistons fans will recall that Chauncey Billups (and Antonio McDyess!) was traded for Allen Iverson at the beginning of the 2008-2009 season. That the timing of my introduction to the advanced stats that praise the Going to Work Pistons coincides with their deconstruction is a cruel irony.
By Wins Produced, Chauncey Billups was one of the best point guards of his generation, and Allen Iverson is one of the most overrated at his position of all time. McDyess was no slouch himself, and ultimately led the 08-09 Pistons in Wins Produced after returning to the team. The decline we have all witnessed was one that Berri’s Wins Produced predicted almost perfectly.
So in a very poignant way, my experience as a Pistons fan — having witnessed the Bad Boys, followed by the Teal Era decline, then the resurgent Going to Work Pistons, followed by the current decline — coupled with my playing experience primed the pump for my interest in stats. I knew that rebounding, defense, and efficient offense won basketball games. I’d been part of teams that did those things. Since a kid, I had watched dominant Pistons do those things.
Reading The Wages of Wins was an “aha” moment for me that resonated with my experience, and motivated me to do a lot of reading and a lot of research.
While I don’t consider myself an expert in statistics, but I do have some training and can understand the work that Berri, (Dean) Oliver, (John) Hollinger, (Wayne) Winston, and others are doing. My own research has led me to the conclusion that Wins Produced is the most reliable all-in-one stat available for evaluating individual players.
Were you skeptical of that type of statistical analysis at all at first? And as a former college athlete, do you understand where the occasional push-back or skepticism against some of the more complicated measures comes from or maybe why athletes are more hesitant to embrace?
Those who know me know that I’m skeptical to a fault about almost everything. Part of my training in college was in Philosophy, which reinforced those inborn traits. Yes, I was skeptical, and yes, I understand the push-back. I get where the skeptics are coming from, because I am one at heart.
One of the things I encourage skeptics of statistical analysis of sports to consider is the relationship between music and mathematics. Obviously, I’m going to oversimplify for the sake of brevity, but an art like music can be expressed in very meaningful ways by numbers. There is an incredibly complex relationship between the two, and learning about that relationship can enhance one’s appreciation for both.
Analogously, basketball is in many ways an art, and I’d venture to say that all basketball fans appreciate the sport for that reason. For example, watching the USA Olympic Basketball teams decimate their competition in style truly is a thing of beauty. When LeBron James is running the break, and executes a perfect no-look alley oop pass to Durant, you can’t help but ooh and ahh.
But at the end of the day, success and failure are expressed by the numbers on the scoreboard, not by how impressive your crossover or step-back jumpshot looks.
What is the key to bridging the gap between so-called believers and non-believers when it comes to statistical analysis? Do you feel that it’s on the advanced stats community itself to not only come up with newer, better measures all the time, but to also do a better job of making complex numerical data more easily understandable? Are there ways to do that even?
Advanced statistical analysis is a relatively new thing to basketball. In baseball, advanced analysis has been around for much longer and has been embraced by many teams, and arguably the majority of them. In many ways, I think time is important. Many of these metrics haven’t had the chance to stand the test of time because they are so new.
That said, I don’t expect that gap to go away in basketball anytime soon, and maybe not ever. There will always be those who’d rather trust their eyes and their gut feelings rather than the numbers.
However, there are ways us “believers” can improve, both in terms of method and in terms of telling our stories.
The stats can keep improving. I find Wins Produced to be a very reliable box-score derived tool, and the box score tells us a lot, and it tells us about the most important things (like number of possessions, scoring efficiency, turnovers, and rebounds). Wins Produced was recently revised to share the credit for a defensive rebound between the individual and the team that helped to create the rebounding opportunity, for example.
However, there are a few things that box score doesn’t tell us. Individual defense is only captured through steals and blocks, for example. It might be useful to know how contesting shots impacts the shooting efficiency of the opponent, but that’s not an easy task. To my knowledge, reliable data related to contested shots isn’t available at the individual level right now, so we can’t do much more than guesstimate. We know that open shots are easier than contested shots, but because we don’t have that data at the individual level, it’s hard to say much more than that.
“Hockey assists” might be another piece of helpful data, but again, it’s not currently available to the public (although reportedly some teams track this individually), so we can’t incorporate it into our models.
That said, I am convinced that the stats are developed enough to say meaningful things about basketball and the productivity of individual players. I think the basketball statistics community as a whole is still learning about the best ways to communicate that information.
For example, if you and I were to sit down and argue about the relative value of Kobe Bryant and Michael Jordan, I would argue that Jordan at his best was a significantly more valuable player than Kobe at his best, which is illustrated by their Wins Produced per 48 minutes in their respective primes.
To a casual basketball fan, that’s jibberish.
However, if I instead explained why Jordan posted those better numbers, i.e., he was a much more complete player who rebounded better, assisted more scoring opportunities for teammates, shot the ball better from two, three, and the free throw line, etc., I might have a little more success, at least in terms of explaining my position.
Seeking out real life, understandable examples is definitely a way we can do better at bridging the gap.
Do you feel like paying more attention to advanced stats could’ve helped you at all as a player? As fans and writers, they’re certainly handy for comparative purposes. But if you were in a coaching or player development position, how would you use statistical analysis as a means for individual improvement?
I am not sure how much it would have helped me as an player, but that’s because I had wonderful coaches who understood what it took to win basketball games. A better question might be how my playing career had been affected had I heeded my coaches advice better! If I had a basketball mulligan, I would work harder on the offensive and defensive glass. I think I had a good understanding of what good shots and bad shots were, and I was a relatively high efficiency guy. However, I didn’t necessarily understand how important rebounds were, and I was fortunate enough to play with guys who did and for a coach that rewarded those guys.
If I were developing young players, there are three lessons I would focus on at the individual level: shot selection, rebounding, and turnovers. Possessions are the currency of basketball, and how a team uses its possessions (and affects the possessions of its opposition) goes a long way toward determining outcomes. These three categories all impact possessions, and all can be affected at the individual level.
Avoiding turning the ball over and/or turning over your opponent helps the team win by creating more possessions than the opponent. In other words, taking care of the ball and forcing turnovers gives you more money to spend.
Out-rebounding your opponent works in much the same way – the more rebounds you get, the more possessions your team has to work with, and everyone, regardless of skill level, can help the team create rebounding opportunities and rebound the ball.
Finally, taking open and high percentage shots increases the chances of your possession paying off. At the individual level, this means developing ways to score near the basket, to draw fouls to get to the free throw line, and/or a three-point shot, while avoiding middle and long range two’s (which are the worst shots in basketball). If you’ve ever during a Pistons game, you’ll know that long two’s are my pet peeve.
It has been my experience that coaches of developing players often scold them for taking three point shots. By the numbers, though, coaches would do better to minimize two point jumpshots and focus on fostering more high percentage opportunities.
That said, I think most of the stats we have now are probably more useful for personel managers than they are for coaches. The all-in-one metrics, like Wins Produced, PER, and Win Shares, give one number that relates the player’s productivity relative to the box score. Which is great, but it doesn’t tell you why that player is productive.
Obviously, the “why” of productivity is important to GMs, but perhaps more so to coaches. You don’t want to put five guys on the floor who are all really good at rebounding and steals but below average at shooting, and so on. Ideally, you want to assemble lineups where strengths and weaknesses complement each other, such as the (Dirk) Nowtizki/(Tyson) Chandler frontcourt that was so instrumental to bringing Dallas the NBA Championship two years ago.
I pay attention to advanced stats (at least the ones I understand) and try and factor in as much available data as possible. But there are still players I love who the numbers hate (Allen Iverson, for example). Are there players who you are able to watch, like and appreciate even though statistically speaking, they’re maybe overrated? Or conversely, who are some players you didn’t pay much attention to but after seeing some deeper statistical analysis came to appreciate more?
I’m with you on all counts here.
Richard Hamilton is a player that I idolized in a lot of ways as a young player, but by advanced metrics, Rip has been a very unremarkable player. My love for Rip’s game stems from my high school coach, who told me to watch two players – Reggie Miller and Richard Hamilton – and learn the ways that they moved without the ball to create scoring opportunities. Patterning my own off-the-ball movement after those guys helped me as a player, and interestingly, it also helped me become a much better screener, as well as cutter. Currently, Derrick Rose is a blast to watch. He’s freakishly athletic, is explosive off the dribble, and finishes in all sorts of creative ways. But by the numbers, he’s nowhere close to a legitimate MVP candidate.
Conversely, I’ve always had a soft spot for players labeled as “hustle” and “blue collar” guys, but metrics like Wins Produced have helped me conceptualize just how valuable those players are. From the Bad Boy era, guys like Laimbeer and Rodman come to mind, and of course Ben Wallace from recent memory. Currently, I’m big on guys like (Kenneth) Faried, (Joakim) Noah, (Serge) Ibaka, guys who make critical contributions to their team but are often overshadowed and under-appreciated. Sticking with point guards, in spite of my rivalry-fueled hate of the Boston Celtics, Rajon Rondo is a brilliant player that helps his team win without scoring lots of points, and thus without as much recognition as he deserves.
- Hoop Convos: Yago Colas
- Hoop Convos: Kirk Goldsberry
- Hoop Convos: Marc Andreas of Net Value