A few years ago, my son joined me to watch a Wizards game. Somehow, he stuck with the broadcast through my frequent pauses to break down plays. I explained sundry topics like why a player praised for being a good defender actually wasn’t. Or where the team’s offensive plays were breaking down. Or how the opponent was targeting weaker defenders by design.
Anyway, Steve Buckhantz was rat-a-tatting stats like: The Wizards have the third most offensive rebounds since January fifth or some such nonsense. WASHINGTON LEADS THE NBA IN THIRD QUARTER SCORING ON TUESDAY NIGHTS WITH A 7:30 TIP-OFF!
After maybe the ninth piece of trivia twaddle, my son asked, “If the Wizards are so great at everything, why do they always lose?”
As it turns out, I’d begun asking a similar question many years previous when I was about the same age. I was a basketball junkie by then. I played as often as I could, and when I wasn’t playing, I was usually finding some hoops to watch. Or I was sketching out my own playbooks. Or digging up coaching journals to study.
For me, the question came when I consumed what passed for analysis on TV and in magazines. The “experts” kept asserting what made teams and players good. Except, their comments often didn’t track with game outcomes. I knew they were wrong, but I didn’t know why.
TSN published raw NBA stats each month. The book got me to launch a rotisserie NBA league with my brother and some friends. It also got me looking at data for the first time and gave me an awareness I could use it to test the veracity of “expert” analysis.
And what I found in the numbers was consistent with what I’d always suspected: the experts were full of crap. They weren’t always wrong, but the errors came at such volume that nearly everything I heard was suspect. So I started doing my own research.
I found likeminded people when Al Gore invented the internet, and I used some of their approaches and invented some of my own. In some cases, I came up with innovative statistical tools that were long-established mathematical formulas. For example, my Consistency Index (a measure of how consistent players are) turned out to be something called coefficient of variation. When I set about calculating consistency, I’d never heard of it.
I’ve stuck with statistical analysis for decades now because those tools and approaches continue to inform and educate me about the game. What I see in the numbers shows up on the floor...and vice versa. Not always, but more than enough to overwhelmingly show their value.
The biggest stat project I’ve conducted, several seasons worth of hand-tracking individual defense, was the result of wanting to understand a data point. In this case, I was curious about why the Wizards’ defense was so much better when Brendan Haywood was on the court when he was so persistently accused of being soft and weak.
The answer? He was long and active defensively — challenging by FAR more shots than anyone else on the roster, including his backup, Etan Thomas. Eddie Jordan, the head coach at the time, praised Thomas for “playing with force.” Meanwhile, Haywood’s and Thomas’s rebounding was nearly identical, Haywood blocked more shots, and the team was much better on defense with Haywood than with Thomas.
Another reason I persist with the stats: they’re great at making small-but-important distinctions. For example, Player A shoots 500-1000 over the course of an 82-game season. Player B takes the same number of shots, but makes one shot fewer per game over the 82-game season. That one shot per game, which won’t be detected by even an expert “eye test,” means 418 makes. That’s 41.8%. The difference of one made shot per game is the difference between good and terrible.
That’s also 164 points on the scoreboard. How significant is that? The Wizards were outscored by 132 points last season.
Are stats everything? Of course not. If they were, I might not watch 200-300 games a season (not counting preseason, international or summer league games).
The answer to my son’s question is simple, by the way. The stats Buckhantz intoned, and Justin Kutcher now bleats, are often basically meaningless because they don’t tell the story of why a team wins or loses.
Good statistical analysis tells that story. That’s what I attempt to do, with varying degrees of success. There’s a reason teams are investing in all kinds of analytics. They’re all trying to get an edge, and they’re finding value in quantifying what was long hidden.