Basic concepts (not math) at the heart of sports analytics

Sports analytics can help us predict how LeBron ages and how KD might perform for the Warriors. Getty Images

What made Bill James, the father of modern sabermetrics, so effective as a writer? There are plenty of plausible answers, from the revolutionary nature of the way James interpreted statistics in the 1980s to his way with words and biting wit.

In the last edition of his "Baseball Abstract" series, however, James offered a different explanation.

"The secret of the success of this series," James wrote, "was that I was dead in the center of the discussion. I was writing about exactly the same issues that everybody else was talking about, only in a different way."

That's sort of how I feel about ESPN's new analytics vertical. If you're interested in viewing sports "a different way," as James put it, this is a new place to find a common style of thinking across different writers and specialties.

At the same time, the topics you'll find here are the same as on the sport index pages elsewhere on ESPN.com, both because this is a home for articles also published there and because we're still trying to answer the same questions as everyone else.

I've written before that what statistical analysts do in sports has as much to do with history as it does with math. When there's a question worth asking (How will the Golden State Warriors perform with Kevin Durant? How will LeBron James age? Were the Philadelphia 76ers wise to tear down their roster?), what we're often doing is using statistics to sift through basketball history in the same way we naturally do in an anecdotal way.

The rationale is not about the numbers per se so much as the ability to avoid some of the perceptual biases that cause our memory to overstate certain examples while ignoring other potentially relevant ones.

Take, for example, projecting the development of players based on similar predecessors. Finding comparisons for young players has been a part of the scouting process as long as it's existed. Alas, even for someone aware of the potential for bias, that process can be unduly influenced by irrelevant factors like the player's race, the hand he uses, or where he happened to go to college. A statistical comparison tool like my SCHOENE projection system doesn't consider any of those things.

Of course, statistics have always played a role in that process.

"We've got this buzzword, 'analytics,'" says Boston Celtics coach Brad Stevens. "Maybe we have more access to data than we had in the past, but certain people have been using statistics and what they can get their hands on forever. It's not necessarily a new concept, right?"

And it's not necessarily more complex, either. Oftentimes, what we call "advanced" statistics are really anything but in the traditional mathematical sense. Metrics like effective field goal percentage and points per possession aren't really any more complex than traditional field goal percentage or per-game stats.

The advancement lies in the fact that these new statistics better reflect what happens on the court. Basketball knowledge, not math knowledge, is required to understand and develop these measures.

And that's why I'd appreciate it if everyone stopped saying that statistical analysts are "smart." It's well-intentioned, certainly, and often accurate -- depending on your definition of the term. But engagement with statistical analysis doesn't mean you're smart. Likewise, those who ignore statistical analysis are not by definition "dumb," and the latter reflects the real problem here.

Talking about how smart statistical analysts are gives the appearance of closing off the field to anyone who doesn't self-identify as smart because they didn't do well in school, or worse yet were told they weren't smart or good at math. And the same goes for anyone who doesn't place a great cultural value on math ability.

Instead, it's better to think of statistical analysis the way we think of any other field of knowledge, one that requires and improves with study. That process is surely easier to navigate if you think about the world in mathematical terms or are familiar with the mathematical field of statistics. However, it doesn't require a diploma from an Ivy League school to begin.

Consider the case of Houston Rockets analyst Ed Kupfer, who didn't finish high school in his native Canada. Kupfer's interest in basketball statistics was an outgrowth of trying to win debates on NBA Usenet discussion forums. That led him to a discussion group about advanced statistics that became the APBRmetrics message board.

There, Kupfer learned about the growing field of basketball statistics with the help of forerunners like Dean Oliver, while also teaching himself the formal mathematical principles of statistics and how to use programs like R in addition to Excel to calculate metrics. Those lessons paid off when Houston Rockets general manager Daryl Morey hired Kupfer as an analyst shortly after joining the Rockets in 2006.

If analytics are a field to be learned, then hopefully this new home is the first step for some, necessary review for others, and reinforcement for those who already consider themselves experts.

Now, if you don't mind, we've got some questions to answer.