Sports

Projecting Impact: Creating a New Method for Predicting Stars in the NBA Draft

By
benpfeifer
Author photo
benpfeifer Sports Editor

Get to know benpfeifer better
Sports Editor

Published2 hours ago on October 02, 2024

Disclosure
We publish independently audited information that meets our strong editorial guidelines. Be aware we may earn a commission if you purchase anything via links on our pages.
Photo via Sports Illustrated, Maria Lysaker-USA TODAY Sports

I’ve spent the last five months building a projectable impact charting system for the NBA Draft. In an attempt to continue improving my own scouting process, I hand-tracked thousands of individual plays to tabulate scores that hopefully project impact to the next levels. Unlike other impact metrics, the goal of my score is to capture whether or not plays a prospect makes indicate success in the future, rather than simply current impact.

Here is the data as it stands today, sorted by their still-unnamed impact score per 40 minutes. I’d love to normalize scores by possession but I unfortunately don’t have access to that data. Before we dive into 2025 draft prospects, let’s explain a bit more about my process of collecting this information and developing my process.

I extrapolated the idea for this project from my in-person scouting process, where I’d often note pluses and minuses when prospects make positive and negative plays. My goal here wasn’t to represent a player’s performance perfectly, but rather, to enhance my memory of players I watched live without the ability to rewind time. 

Cognitive Biases

Humans can’t escape cognitive biases that confound our thinking and judgment. It’s impossible to fully eliminate bias from draft scouting, but this project attempts to lessen my biases as much as possible. Here are the primary biases this impact-tracking project attempts to lessen:

— Recency Bias, or the idea that we tend to favor recent events and memories rather than older ones. It’s easy to attach our idea of a player to the game/piece of information we most recently encountered. All of the impact scores exist regardless of when the games took place, though we can use the data to track improvement over time.

— Anchoring Bias, or our tendency to favor information we receive early in a decision-making process. It can be difficult to shake your first impression of a player, whether positive or negative. Quantifying the scouting process helps eliminate this time bias, similar to the recency bias point.

— Aesthetic Bias, or our tendency to favor things we perceive as aesthetically pleasing. This is especially prevalent in scouting and a reason why some can overvalue inefficient shotmakers and undervalue quiet production. Take Keegan Murray for example, a prospect who I faded too hard because I overvalued aesthetics and undervalued his raw scoring production.

— Hot-Cold Empathy Gap, or our tendency to underestimate the impact of emotional states on decision-making and thought processes. Underlying moods and emotional states can blur how we perceive certain players. Tracking and quantifying individual plays helps remain grounded, level-headed and present while scouting.

Explaining the Scoring

What does the impact “score” actually mean? The easiest way for me to conceptualize the scores is as a net total of a player’s positive and negative moments on the court. I’ll think about what a play suggests (or doesn’t suggest) about a player’s ability to translate to the next level in some way. I score plays on a scale from 0-2, with zeros being neutral plays, ones being solid positives and twos being strong positives. 

The easiest way to explain this will be through examples. Take this Nolan Traore play for example.

Traore burns his defender off of the dribble and wraps this pass around for an easy layup attempt. The initial advantage creation with Traore’s elite acceleration earns him a point given the value of advantage-creating guards. The pass earns him another point due to the difficulty of the delivery and the size of the window, making this a +2 play for Traore.

Let’s look at this Traore drive:

Traore gets downhill here for an easy bucket. However, Traore benefits from two excellent screens on this stack action and poor defensive play. Though Traore does flash his speed, these kind of plays don’t translate, earning Traore no points.

Here’s another Nolan Traore play:

The defense sags off of Traore due to his lack of shooting threat and he’s unable to create an advantage, earning Traore a -1 here. To be clear, the minus isn’t for the missed shot, but for Traore’s lack of a shooting threat, which diminishes his advantage-creation potential.

Another Traore play:

Traore turns the corner on the hard hedge and forces a rotation, earning him a +1 for the advantage created. He’s unable to decelerate though and runs into a charge, earning him a -1 as Traore’s struggles slowing down are concerning for his driving translation. In the end, this clip totals to a net of zero points.

 

Let’s look at one more Traore play: 

Traore sinks off of the far corner to deny a pass to the role before sprinting out to contest a jumper. Even though this shot falls, Traore earns a +1 here for his incredible closing speed, suggesting his potential as a ground coverage defender off of the ball.

Another way to conceptualize this score is as a heuristic for dominance. We often underrate prospects who consistently imprint themselves on games in a multitude of ways. This play-to-play dominance translates to many cases and the impact projection attempts to consider that fact.

This project isn’t meant to be scientific or a replacement for traditional impact metrics for scouting the NBA Draft. Though I do my best to account for context, role, opponent quality, game situation and other confounding factors, there will be human error involved. To this point, I’ve tracked and scored nearly 7,000 individual prospect minutes for the 2025 class. There are bound to be mistakes throughout as I keep refining my process.

I’m excited to find out how my scores correlate with eventual success in the pros. If I had infinite time and resources, I’d love to backfill games from old draft classes to learn more about how the scores correlate with NBA impact. That’s a project for another day, though.

2025 Draft Scores

Let’s return to the scores for the 2025 draft and what we can learn from them:

Predictably, Cooper Flagg tops the charts and it’s not particularly close. This makes sense, as he’s constantly impacting winning at a high level on both ends with high-level, NBA flashes throughout. He’s the best defender I’ve ever scouted at the high school level and the scoring reflects that. Flagg is incredibly consistent, as his lowest score in the nine games I’ve tracked for him was a +12, which is still an extremely good number. These scores can help determine projectable impact even in games where box score production looks pedestrian.

Age is another important component here, as it’s understandable for younger players to consistently impact games in a way that projects forward. Prospects like Ace Bailey, Drake Powell, Derrion Reid and Hugo Gonzalez don’t score particularly well, but all have flashed high-level ability at some levels and can learn to better apply their excellent physical and athletic traits. Conversely, high-scoring young prospects like Flagg, Ben Saraf, Rocco Zikarsky, Nolan Traore and Dylan Harper could be encouraging for future projections.

More concerning are the older prospects who struggle to make positive plays consistently like Coleman Hawkins, Zeke Mayo and Payton Sandfort. These might be excellent college players, but it’s harder to project 22+ year-olds to improve their ability to project to the next level than it is for younger prospects.

I spent far too many hours rigorously watching prospect film and refining my process here and will continue changing and refining things in the coming year. This is simply a starting point for a project that brings me joy and helps me scout more deeply and intentionally.