Of Numbers and the Importance of Context: A Twitter Rant

This post was inspired by an attempt to watch a Google+ Hangout featuring a number of young basketball bloggers. Granted, for the most part, they’re very young, but much of what was discussed in the few minutes I was able to make it through echoed the feelings of their older peers, whom I know they look up to. Below is a summation of my thoughts on the direction basketball analysis is headed, and why I don’t like it:

I try to watch these basketball discussion hangouts, and it’s just a bunch of reading numbers off a screen. I don’t get the point. The numbers are publicly available. What people aren’t doing with them very well is translating them into actual useful information.

Ex: Jrue Holiday has a higher AST% than Steph Curry, but what does that really mean? It certainly doesn’t mean that he’s a better passer. He could, in fact, be a better passer (he probably is), but that’s central to HOW MUCH he’s passing, not how good he is at it.

When you’re comparing different players in different situations, those numbers mean so little. Discernible skills are far more important. Yet, here people are, discussing at length the results of far different situations and using them to define which is the better player. I’ll try not to go on too much about how that’s distinctly different from the common idea I hear about “the process,” but it makes no sense.

Another Ex: TS% as a measure for better scorer/shooter. Higher is better, yes, but what leads to those numbers is so much more important. Guards who play with true low-post threats always suffer there for the simple fact that the easiest shots are at the rim. Any paint area that’s clogged with bigs is not really the place for a guard to be going all the time, for a few reasons. It hurts the post-player’s positioning and also brings a 2nd man to where he is trying to post up. Also forces them to move. A drive/cut through the lane means they have to wait longer or even give up position they’ve already established. You don’t want that. Lots of guys who are good at penetrating and scoring at the rim sacrifice some of their easier attempts to keep the offense balanced. If you’re not Stephen Curry (better from 3 than even at the rim), then that’s going to hurt your TS% a whole lot.

It just seems silly to try to even use circumstantial numbers, even if partly based on skills, to decide the better player. Skills aren’t circumstantial. They’re real, and though they are not as easily measured, they are far better for trying to judge who’s best. There is no stat for the best ball-handler, the best passer, the most creative thinker. All-encompassing stats try to measure, but fail. It’s as if we’re just throwing out the visual and psychological aspects of the sport and running with numbers and sociological measures. Not everything can be measured in a simple value, and most values don’t allow you to take anything new that can’t be learned visually. Shots aren’t missed due to random variation. Over enough time, what they do on average will work itself out, but there’s reasons they missed: poor form, taking a more-difficult-than-usual shot, better D, fatigue. Under ideal circumstances, a player could defy that variation easily.

I guess I’ve just had enough of people trying to determine any type of causation from stats. It’s a pointless endeavor. It’s taking over basketball conversations. It bothers me to no end that a beautiful game is being turned into something so…bland. Anyone who disagrees with the inferred meaning of said stat is made out to be some sort of uneducated, unlearned fan. The message from the stats community is one of aggressive reform. “There is no going back.” Sometimes, progression isn’t what’s best. Sometimes, things are the way they are because they should be that way. The way I worded that is intentionally confusing. Just because something can get muddled, uncertain, and confusing, doesn’t mean it’s the wrong way. I contradict myself all the time. But there’s nothing wrong with that. With a game as in-depth and complicated as basketball, things are going to get confusing. We don’t necessarily need cold hard facts to solve everything for us. And that’s where the use of stats goes too far.

They’re GREAT at informing us about things we didn’t know about. I love that part of stats. So interesting the stories they can tell. Can help us track changes over time or just from the norm in a single game. They organize knowledge that can be hard to keep exact track of. BUT, they also lead us down a path of trying to explain things with them, without actually watching the game.

My FAVORITE Ex: A volume shooter takes a ton of shots at an average to poor clip. Team loses. People who read the box score: “BALL HOG!” This especially applies to Russell Westbrook right now. People look at his FGA and TS% vs Durant and immediately identify him as one. They don’t think about how, as great as KD is, he’s often underwhelming and that when played physical, can disappear. Neither is true of RW. When RW is in a game, he always puts his mark on it. He doesn’t back down, and is always aggressive, even in low FGA nights.Sometimes, it floats the team. Sometimes, it sinks it. But what it always does is give KD the breathing room he needs to be great. Russell’s aggressiveness allows Durant’s greatest weakness (his passivity) to exist without significant detriment to OKC. Russ takes most of the blame for the team struggles. I understand why. Even visually, the “no-chill” attitude can be tiring. But it’s also clear that if he didn’t shoot as much, Durant wouldn’t be nearly as efficient. Life for KD without RW would be more difficult.

That’s something that none of those numbers can encompass. For everything that happens on the court, there’s a trade off somewhere. Even stats can tell us this. Graph’s of the relationship between USG% and TS% can help us understand the importance of context. For every extra shot you take, the more likely is the defense to plan for it. More defense means less efficient production. Sometimes, even for efficient players, there just aren’t any more shots to be had, due to a limited skill-set.

Tyson Chandler is a great example of this. He’s arguably the most efficient scorer in the league, but shoots just 6 times a game. Often, I see NYK fans calling for him to get the ball more. But to do what? He scores primarily off put-backs and PNR to the basket. He isn’t a good dribbler, shooter, or post-player. He maximizes his skill-set, but you can’t exactly get him more shots. He can only take what the defense gives. Trying to increase his attempts would likely only result in lowering his efficiency numbers.

Another player, James Harden, was one of the most efficient players in the league last season. This year, he’s far less so. Obviously, that’s because his role has increased dramatically, both in minutes and in the skills he’s asked to showcase. He also no longer plays alongside All-Stars Durant and Westbrook. Basically, the point is, every environmental change impacts numbers.

Therefore, those numbers are in large part a product of the environment, likely as much as they are the player. Trying to gauge similar players off of what the numbers say will get you nowhere but back to the results you’re already judging.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s