Revisions, Pt. I

Posted by Rob Mahoney on August 10, 2010 under Commentary |

Last week’s foray into the positional revolution was a good start, but there’s still a lot of work to be done. As I noted previously, Drew Cannon’s positional system isn’t coming out of the internet womb fully-formed; a lot of adjustments and tweaks are necessary for the model to become viable. Still, Cannon’s design offers a welcome starting point for both discourse regarding positional fluidity and, hopefully, some eventual long-term change in the way we think about and define positions.

There is no end to this process. Even if we successfully shed the five traditional positions in favor of some other system, players and their roles will continue to evolve. It’s critical that we’re constantly challenging the limits of positionality to match with the on-court product. Note that those limits aren’t being tested without reason. It’s important that positional rhetoric remains descriptivist in nature. We’re not saying “this is the way that position X should play,” but rather “this is the way that position X does play.”

With those things in mind, any model proposed here needs to be poked and prodded. Finding the leaks is an important step in the process. I’ll be the first to concede that no positional model will ever be perfect, but tinkering with the system’s imperfections is the the path that will bring us all closer to that flawless standard.

~~~~~

Thus far, Cannon’s defensive positions seem to be the source of the most controversy. While there is some confusion over what Cannon’s designations do and should mean, the most glaring problem was captured perfectly by Tom Ziller:

The problem with the way this new development is framed is that it still relies on demonstrably imprecise labels. If Rodrigue Beaubois is a “D1″ — meaning he guards point guards despite often playing shooting guard next to Jason Kidd or J.J. Barea — then you’re assuming there are “1s” for him to guard, which is just the type of assumption the Positional Revolution aims to destroy.

The point of a new nomenclature is to do a better job describing what players actually do, and do well. If “1s” no longer exist, in favor of the Cannon/Mahoney use of “scorer” or “creator” for offensive roles, what does a “D1″ do? And is that by necessity (due to size or athleticism) or ability? Is calling Beaubois a “D1 Scorer” any different from calling him a “guard”? A more useful classification might be something like “DPick-and-RollA+” or “DPostC-”.

In their original context, defensive designations like D1, etc. are actually very counterproductive. If the hope is to move away from traditional positions, it makes little sense to lean so heavily on them.

Yet Ziller’s suggested variant, while useful, is both highly subjective and goes well beyond what position is thought to bring to the NBA discussion. There are some skill valuations inherent to positional delineation, but in my eyes (and this point is certainly up for debate), it’s important that positions describe and group without necessarily assessing how well a certain player executes their role. Those types of appraisals require far more nuance, and using them to define players via position sacrifices the accessibility of those classifications.

In addition, the wide variety of defensive abilities each NBA player is asked to display makes it difficult to pin down a primary role via skill set. For purposes of convenience, the defensive positions should be as short as possible, so saying that Dwight Howard is a “DPick-and-RollB+/DPostA/DHelpA+/DHelpPostB/DRebounderA+,” while descriptive, is probably a bit silly. Yet, if we’re going to classify players by their skill (and skill level, in this scenario), Howard’s post defense is no more important than any of those other facets. How would one accurately and succinctly convey all of that information in a usable (and more importantly, re-usable) manner?

I think the key is to step away from skill descriptors, particularly at such depth. While it makes sense to describe a player’s offensive role as a “scorer” or “rebounder,” defense functions much differently. A player’s defensive position (not utility, position) hinges on, again, not how well a player defends, but what purpose they serve. Or, for simplicity’s sake, what types of players they’re able to defend.

Who a given player is able to defend certainly ties into the respective skills of both the defender and their offensive counterpart. At a more basic level though, a player’s defensive range is determined by the various heights and speeds he’s able to counter. I’m not talking about the height and speed of the defender, but rather, the ranges of those two variables that a defender is physically able to contest.

Thus, I offer the following modification of Cannon’s defensive positions:

We’ll use the original D1, D2, D3, etc. designations, but with each describing a certain range of relative size and speed. D1 no longer represents a player’s ability to defend point guards per se, but their ability to defend shorter, quicker opponents. Likewise, a D5 would indicate one’s ability to defend bigger, slower players, regardless of one’s own size and speed (Chuck Hayes, for example, is 6′6”, but would be a D5 because of his ability to guard taller, stronger players).

A few notes:

• With this system, we eliminate the idea that there are 1s or 2s to guard, and instead simply assume that there will be players of different sizes on the court. Someone will need to guard them. Teams don’t have to match small with small, but they do need to match a small opponent with a D1 (even if that D1 would be a shooting guard or small forward by most conventional standards).
• It’s important that the boundaries between the size/speed of D1 vs. D2, D2 vs. D3, etc. are nonspecific. This is not meant to be measurable, as creating the necessary framework would involve drawing far too many arbitrary brightlines.
• Positions are meant to be convenient. As such, we really do need to sacrifice some depth for the sake of easier use and better understanding. The point isn’t to create some undecipherable code that no other NBA fans can solve, but to create a relatively uniform system that’s a bit more descriptive and accurate than the current one.
• These positions are different from traditional ones, particularly because they account for defensive versatility and cross-matching. For a basic example, let’s take Rodrigue Beaubois. He may start at shooting guard this year, but in the Allen Iverson mold, will mostly be cross-matched on opposing point guards. So from a positional standpoint, it makes far more sense to call him a “Scorer, Creator/Handler, D1″ than it does a shooting guard. It gives us a bit of insight into what kind of players Beaubois will be guarding, as well as his offensive responsibilities.
• For a more complicated example, look at LeBron James. He’ll be listed as a small forward, but we know his offensive role is more far-reaching than the limits of a traditional wing. Additionally, LeBron has become such a useful defender that he can guard all kinds of positions. Thanks to his incredible combination or size, speed, and strength, one could make a legitimate argument that James is actually one of the few players capable of defending all five traditional positions. His ‘SF’ label has never done him justice on offense, and now it’s just as constricting defensively.
• There’s still something to be said about how a player defends that’s completely unaccounted for. It’s distinct enough from the question of ‘How well?’ that it could technically be incorporated, but I see no simple way to incorporate it. Something to put on the wish list, for sure, but at this stage those distinctions seem a bit too complex.
• Another concern is addressing players who can’t really defend anybody. Regardless of where we put the bounds of a positional system, there are going to be exceptions. There will always be someone who doesn’t fit neatly into the given categories. Tentatively, these players will be addressed as ‘D0,’ but it’s certainly an idea worth revisiting.
• Please, leave questions and concerns over this system or propositions for other defensive positional models in the comments. Feedback is a crucial part of this process, and every reader is an invaluable part of the refinement of this system.

Special thanks to M. Haubs, Matt Moore, and Zach Harper for spiritual guidance.