Structure and Event

 Player analysis is hard. We all know it’s true. It’s the reason why two people can watch the same player, possession or stat-line and walk away with two vastly different conclusions on whether they’re actually good or not.

If this issue were yet another symptom of our decaying online discourse, then there’d be no use in me writing this article or you reading this article. However, this same uncertainty around determining and quantifying player impact still persists at the highest levels of our sport, and it’s the reason why institutions, armed with premium datasets and hordes of full-time scouts, still get it wrong. Every year, draft boards fracture, bad contracts are given out like candy, and team-building philosophies collapse under any sort of playoff inspection.

Take the years-long debate over who should be the number one pick in this year’s upcoming NBA Draft out of Cameron Boozer, AJ Dybantsa, and Darryn Peterson. For some, Cameron Boozer is the drop-dead choice. A player who represents the Platonic ideal of prospect productivity, currently in the midst of one of the greatest freshman seasons of all time. At the same time, however, you have anonymous NBA scouts questioning whether he’s even a franchise cornerstone at the levels of his contemporaries.

What these issues and debates reveal is not confusion of talent or disagreement over players, but a deeper and more intrinsic struggle over what exactly player impact is, and how it should be determined. To me, at least, the issue stems from how we treat player impact. As if it were a single, unified concept that can be ranked from good to bad. So, before we ask the question of ‘who’s better than who’ or ‘who deserves what contract’, we need to first define: what exactly is player impact?

The Two Degrees of Impact

The framework I will try to explain in this article stems from a simple but too often overlooked truth about basketball, which is that not every action carries the same value. And over the course of a 48-minute game, hundreds and hundreds of actions will occur with varying levels of influence.

This is because basketball (and many other team sports) is, at its core, a stochastic sport. This means outcomes emerge from a web of probabilistic events rather than deterministic sequences, and small variations compound across possessions. As a result, actions cannot be treated as equal simply because they occur within the same possession or box score.

Every action that takes place on a basketball court is a probabilistic bet that seeks to tilt the game-state towards one direction; however, most mainstream analysis seeks to collapse all of these actions into narrow sets of outputs: points, rebounds, assists, etc. And in doing so we end up over/under valuing certain player archetypes that better suit our own personal biases.

Cumulative Impact

Following on from that jargon, we can now get into the weeds of the framework, and the first idea I need to introduce is that there are two levels of impact that a player can exert on a game, with the first being what I currently refer to as cumulative impact. Cumulative impact is the influence a player exerts on a game through actions whose value emerges from repetition and continuity. Cumulative players generate and preserve extra possessions, and while these actions do not produce high-value outcomes in isolation, they shape the conditions in which future possessions occur. Rather than deciding outcomes outright, these actions reduce the team’s reliance on low-probability solutions.

Basketball behaviours that indicate a player is cumulative, and some statistical indicators:

ability to extend and complete advantages

• assist rate 
• potential assists
• secondary assists 

ability to generate extra possessions

• offensive rebound rate
• steal rate
• offensive fouls drawn
• deflections

ability to preserve possession integrity

• turnovers per game 
• turnover percentage
• assist to turnover ratio

ability to impact possessions without direct ball involvement

• gravity rate
• opponent rim attempt frequency
• opponent fg% at rim
• defensive on/off

However, while cumulative impact explains how teams build and sustain momentum, it doesn’t explain how they convert that control into outcomes. You can control the chessboard for hours, accumulate pieces, dictate the pace of play, but at the end of the day, the game is still decided by checkmate.

Decisive Impact

This brings us to the second level of player impact, which is what I’m calling decisive impact. Decisive impact is the influence a player exerts on a game through high-difficulty and high-leverage actions that directly resolve possessions and sharply alter the game state. Where cumulative impact shapes the flow of a game, decisive impact collapses that flow into results. These moments disproportionately affect win probability and rely more on individual execution under pressure.

Here are some behaviours that indicate a player is decisive, and some usual statistical indicators:

ability to independently create/prevent advantages:

• pts/75
• ts added per 100
• shot usage
• FTA per 100
• block rate

ability to execute when time, space or structure is constrained:

• grenade percent
• grenade efg
• isolation efficiency
• clutch efficiency
• contested shot efficiency
• STOP%
• rSTOP%

The Four Quadrants of Player Impact

I’ve always been uncomfortable with binary thinking, and it’s probably why, growing up in school, I was always drawn more towards the humanities than STEM subjects. To me, very few meaningful things in life fit neatly into two categories, and player impact is no different.

In reality, we know that players vary in just how much cumulative or decisive impact they exert on games and that these dimensions are independent of each other. It’s why a player can be highly cumulative without being highly decisive, and why a player can be highly decisive without being highly cumulative.

Conceptually, I think this framework is best visualised as a fluid, four-quadrant model:

Low-Impact Players

These are the players whose presence does little to meaningfully alter either the structure or outcome of games. They neither accumulate influence through their repeated activity on the court, nor do they have the capability to resolve high-leverage possessions. This does not mean they lack basketball skill, but rather that their actions do not scale in either volume or difficulty. YOU SHOULD AVOID DEDICATING EXTERNAL RESOURCES TO THESE PLAYERS AT ALL COSTS.

Structural Players

Players in this quadrant are your textbook floor-raisers. The RAPM darlings. Rather than decide games outright, they shape how games are played through influencing a large volume of possessions. Their impact is rarely spectacular in isolation, but compounds over time as they increase the frequency of point-scoring opportunities. YOU SHOULD DEDICATE RESOURCES TO THESE PLAYERS PROVIDED YOU HAVE THE REQUIRED EVENT PLAYERS TO CAPITALISE ON THEIR IMPACT.

Event Players

Players in this quadrant are your archetypal ceiling raisers. They are those players whose value lies in their proficiency in high-difficulty situations. The Ball Don’t Stop alumni, if you will. And while their impact might not be felt continuously over a game, it will no doubt manifest itself in those high-leverage moments where only the strong survive. YOU SHOULD DEDICATE RESOURCES TO THESE PLAYERS PROVIDED YOU HAVE THE REQUIRED STRUCTURAL PLAYERS TO PROVIDE THEM AMPLE CHANCES FOR IMPACT.

Dual-Impact Players

This is the quadrant occupied by the true outliers of the sport, and those special hoopers who had/have the ability to both: increase the amount of and resolve their team’s scoring possessions. There is rarely a moment in a game where they are not either directly or indirectly involved in a possession. YOU SHOULD DEDICATE THE MAJORITY OF YOUR RESOURCES TOWARDS TRYING TO ATTAIN THESE PLAYERS.

Why the Most Valuable Players Tend to be Dual-Impact

An important, and perhaps the most important clarification that needs to be made before going further, is that within the same quadrant/archetype, players can vary enormously in quality. Not all structural players are equally valuable, nor are all event players, and this framework, when applied lazily, can flatten those distinctions. To me, what explains these gaps in quality are not the archetype itself, but the constraints under which players operate. Inherent physical and biomechanical limitations, as well as the external team environment, shape the range of total actions a player can perceive as viable. And while you can adapt your game around these constraints to become an effective player, these constraints will always exist and limit your ceiling.

A useful way of conceptualising this is through Plato’s Theory of the Forms. In Plato’s theory, the physical world contains imperfect instantiations of ideal Forms, abstract, complete expressions of concepts like justice or beauty. A drawn circle may resemble the Form of a circle, but it will never fully realise its perfection. Dual-Impact players represent the ideal Form of player impact: the fullest and most complete expression of game influence, where the physical capacity, technical skills, and cognitive processing all align. Structural and event players are merely imperfect (but still valuable) instantiations of that form, approximating along one or two of those dimensions, but not all. And it is this incompleteness that introduces fragility into their impact when conditions change.

Beyond any specific technical or physical traits, what truly separates dual-impact players from the rest of the field begins in the mind. At the highest levels of the sport, where technical skill-sets overlap and physical margins narrow, what separates players is not simply what actions they can execute or how they execute, but when and why they execute. Dual-impact players show a heightened sensitivity to the spatiotemporal structure of the game and understand that possessions are not isolated events, but linked sequences whose value compounds across time. This cognitive elasticity is what allows them to constantly shift their thinking from an individual to a team level, and the emergent result of this advantage is that their individual impact is less sensitive to their external environment.

Event and structural players, by contrast, are constrained by the cognitive tax of over-specialisation. Event players are forced to operate in high-leverage moments repeatedly, becoming over-reliant on situations that cannot be sustainably generated, and structural players become over-reliant on teammate continuity and control. In both cases, impact becomes brittle and increasingly sensitive to the external environment.

Basketball, ultimately, is a team sport, and any individual player’s impact only exists in relation to the environment it shapes. This is where the value of dual-impact players becomes overwhelming. Because they influence both the conditions under which possessions occur and the outcomes those possessions produce, they reduce the sensitivity of a team’s success to lineup construction. This is because it allows their GMs and coaches to allocate resources more efficiently, to surround them with lineups that lean towards offense or defense without sacrificing anything. This is why historically great teams like the OKC Thunder now or the Golden State Warriors in the past could field such defensively slanted teams around Stephen Curry or Shai Gilgeous Alexander, as the multiplicity of their skill-sets means they could make up for any gaps in offense or defense.

So…Why Do We Keep Getting It Wrong?

If dual-impact players represent the fullest expression of influence within a basketball game, then it raises an uncomfortable question: why are they so often misidentified and misunderstood? At a fundamental level, we do not evaluate impact directly, we evaluate signals of impact. Basketball, as both a sport and spectacle, privileges actions that are visible and narratively decisive. Event impact fits nicely into this perception, and a tough iso bucket or a chasedown block are immediately legible to both audiences and decision makers. As a result, event players are intuitively perceived as controlling games even when their impact is episodic rather than systemic. This perceptual bias is reinforced by the inherent structure of basketball discourse itself. Highlight clips, box score, scoring averages, and usage rates all reward possession-ending actions, and even most advanced metrics remain possession-local in nature. The consequence of this is that players who exert influence by shaping the conditions of future possessions (through indirect actions) struggle to announce their own value with the same clarity.

Due to this, event players are often easily mistaken for dual-impact players. Their ability to generate high-leverage outcomes under pressure is rightly valued, but when this capacity isn’t paired with the ability to frequently improve the frequency and quality of possessions, then its overall influence is too limited. Crucially, the inverse error also exists, just in fewer cases. Structural players are often overvalued in environments that reward stability and scale, where their ability to raise baseline efficiency can be mistaken for universal sufficiency. In this sense, both event and structural players are epistemically misleading, just in inverse ways. Event players are often overvalued due to the visibility of their impact, which makes it easy to narrativise, and structural players can be overvalued for those same opposite reasons.

There is also a further distortion that emerges online in spaces where we debate basketball, once this perceptual imbalance is recognised. Because event impact is more immediately legible and thus accessible to casual fans and mainstream media discourse, more analytically literate observers can start to define their analysis in opposition to it. Structural impact, requiring specialised language and a trained eye to identify, becomes not just a category of value but a signal of understanding itself. In this way, otherwise intelligent analysts can reproduce the same error as a casual observer, converging on a form of groupthink where what is hardest to narrate is assumed to be most important.

Implications for the Draft and Free Agency

Once player impact is understood as context-dependent rather than absolute, the implications for roster construction begin to clarify themselves a little more, especially when it comes to the draft and free agency. This is due to the fact that both of these mechanisms operate under different constraints and risk profiles, which change how different impact archetypes are valued.

Let’s start with the NBA Draft, which is by definition an exercise in projection under uncertainty. Teams that are rewarded with a lottery pick are generally weaker teams that lack organisational coherence and are years away from playing meaningful playoff basketball. In these developmental contexts, event impact is inherently fragile. High-leverage, possession-deciding actions only accrue value when a team is already capable of consistently generating competitive possessions. Without that foundation, event-oriented prospects are forced into roles that exaggerate their weaknesses and develop bad habits despite gaudy box score numbers or Twitter mixtapes. We also must acknowledge historical trends and how raw scoring output has lost a lot of its relative value due to the skyrocketing of league-wide offensive efficiency. Kobe Bryant’s 35.4 points per game in 2005-06 came in a league environment where the league average ORTG was 106.2, and teams averaged roughly 94-96 possessions per game. Luka Doncic’s 33 points per game this season, however, is occurring in an environment where league ORTG exceeds 115 with roughly 101 possessions per game. This means that as the entire league has become more efficient, the marginal value of raw volume scoring has declined. This makes structurally-inclined prospects, those who generate extra possessions and stabilise lineups, the more robust draft investments, even if their high-end outcomes don’t appear as high as the next prospect.

Free agency, however, operates under some slightly different logic. Established teams, those built around one or more dual-impact players, already possess the structural capacity to survive the regular season and get to the high-leverage environments of the playoffs. In those contexts, the marginal value of additional cumulative impact diminishes, while the value of decisive impact increases. It’s in these arenas where event-oriented players who struggle to justify primary roles can truly shine, as their ability to resolve possessions can meaningfully swing playoff games. When used in shorter stints or in tilted matchups, they can buoy the defects that come with their profiles over large sample sizes.

This asymmetry helps explain why certain player profiles appear to be ‘overdrafted’ or ‘underpaid’ early, only to resurface as valuable contributors later in their careers. (I’m looking at you, Aaron Gordon and Andrew Wiggins). The most shrewd organisations understand this distinction and resist the temptation to evaluate players as static entities.

However, there is a human element to these decisions that prevents teams from disavowing the historical mistakes of the past. Draft choices must be justified immediately, and event-heavy profiles must provide a form of narrative insurance. Their impact is legible, and their upside can be sold easily to owners and fans. “The talent was undeniable” is an excuse that allows front offices to skirt public scrutiny and keep their cushy, million-dollar jobs in NBA executive roles.

Taking a Wider Look

A framework is only as useful as its ability to scale, and the distinction between cumulative and decisive impact does not stop at the individual level. In basketball, where five players share the floor and possessions are both numerous and interconnected, team performance is best understood as the aggregation of individual impact profiles. For all the Xs and Os and schemes a coach wants to run, you can’t coach out a player’s built-in tendencies they’ve developed over years and years of reps.

At this point in the article, I think it’s clear that what makes dual-impact players so valuable is the fact that they don’t grow on trees. And so, a lot of what team-building becomes is being elite at aggregation and figuring out how to maximise interactions between incomplete skill-sets. Procure too many event players on one team, and you become over-reliant on situations you can’t sustainably generate. Your success as a team hinges on shot-making and individual brilliance, which causes you to become a high-variance team. However, if you acquire too many structural players and you become over-reliant on possession and territorial control, which causes you to become a low-variance team. We can observe that in both of these cases, stacking too many of one impact type on your team causes you to increase your margin of error, which matters all too much in a large sample size sport like basketball, where outcomes are decided across multiple games across a season.

This tension helps to explain why regular season and playoff basketball often reward different team profiles. This balance between cumulative and decisive impact begins to explain why certain teams consistently out/under perform expectations in specific competition formats. In large sample size formats such as the regular season, teams built with strong cumulative foundations tend to be rewarded for their efforts. Their advantages lie in reducing variance in possession quality. Over an 82-game season, those small, repeated gains compound, insulating them from the inevitable shooting slumps that derail more volatile teams. However, as formats compress, be it a playoff series or a single elimination cup game, the equilibrium begins to shift. Smaller samples magnify variance and reduce the value of pure accumulation. In those environments, teams with a greater share of decisive impact are better equipped to survive volatility, as they possess players capable of converting limited opportunities and swinging games on only a handful of possessions.

Ultimately, the same distinction that separates players at a micro level re-emerges at the macro level of teams, which is the cognitive difference. Teams built around event-heavy profiles demand a high level of mental bandwidth to constantly pull a rabbit out of the hat to win games, and, conversely, teams built around structural players demand a level of constant concentration and collective synchronisation over long stretches, to a point that isn’t humanly possible. What separates the best teams, much like the best players, is their elasticity and adaptability, and there is an equal cost of leaning too far into one identity, just in a different currency.