How to Make Winning NBA Over/Under Picks for the Upcoming Season

As someone who has spent years analyzing sports data and, perhaps just as importantly, watching countless games not just as a fan but as a professional observer, I’ve come to view the NBA season as a dynamic, ever-shifting landscape. Making winning over/under picks isn't about finding a static number and locking it in; it’s about navigating a track that can warp and change beneath your feet. This reminds me of a concept from a completely different arena—video game design. I recently read about a racing game where the track dynamically shifts, warping you to unpredictable environments. You can't just memorize the curves; you have to stay alert, adaptable, and ready for a tight-turn candyland, a bouncy mushroom forest, or an airborne stunt show at any moment. That’s the perfect metaphor for the NBA season. You might think you have a team’s trajectory memorized—their roster, their coach’s system, their historical pace—but then, before you know it, a key injury, a surprise trade, or a sudden coaching change warps the entire context. The general outlines of the league’s “worlds” are knowable, but never knowing exactly which sequence is coming is what makes the analysis both maddening and thrilling.

So, how do you build a framework that respects this inherent unpredictability? The first step is to abandon the idea of a single, definitive number. The sportsbooks set the line, often with remarkable accuracy, based on a mountain of data. Your job isn’t to out-compute them in a vacuum; it’s to identify where the collective market perception might be warping reality. Let’s take a concrete example from last season. Many models, including my own early projections, had the Memphis Grizzlies pegged for around 48.5 wins. The general outline was clear: a young, athletic core with a defensive identity. But the season became a brutal lesson in dynamic shifting. The Ja Morant suspension was the first warp, tossing us into a different tactical realm. Then, the catastrophic injury to Steven Adams, followed by a cascade of others, effectively transported the team to a “bouncy mushroom forest” of lineups they never planned to use. The final win total landed in the 20s, a staggering deviation. The lesson? The base projection is just the starting track. You must assign probabilities—not just guesses, but quantified estimates—to potential warp events. What’s the likelihood of a 20-game injury to a top-three player? For a team with aging stars, I might assign a 35% chance. For a younger squad, perhaps 15%. This isn’t pseudo-science; it’s about explicitly accounting for variance.

My process always involves a deep dive into what I call “pace and space” catalysts, which are the primary drivers that can trigger those seismic warps. Coaching changes are a huge one. A team moving from a slow, grind-it-out coach to someone like Mike D’Antoni isn’t just changing plays; they’re changing the fundamental physics of their games. The over/under for total points in their games might need a 5 to 7-point adjustment overnight. Player development trajectories are another. Look at the Oklahoma City Thunder last year. The consensus win total was around 23.5. But if you watched Shai Gilgeous-Alexander’s closing stretch the prior season and saw the arrival of Jalen Williams, you might have sensed a warp into a higher-scoring, more competitive “world.” They smashed that total. I had them at 38 wins in my model, which felt aggressive at the time, and they still exceeded it. Sometimes, the data points to a shift before the market fully prices it in.

Then there’s the art of synthesizing the numbers with the narrative, and this is where personal perspective is non-negotiable. Analytics might tell you a team’s net rating suggested they “should” have won 45 games, but they only won 40. The market often leans on this for the next season, setting an over/under at 44.5, expecting positive regression. But you have to ask: why did they underperform? Was it bad luck in close games—a stat that does tend to regress—or was it a deeper flaw in clutch execution, a soft defense, or a lack of leadership? I remember evaluating the Chicago Bulls a couple of seasons back. The numbers screamed “regression candidate!” after a poor record in close games. But watching them, I saw a team with a stagnant half-court offense and defensive lapses that weren’t fluky; they were systemic. I leaned under, and it was the right call. The fuzzy, visually rough transition in that racing game? That’s the uncomfortable gap between the clean projection and the messy reality of human performance. You have to squint through that fuzziness and trust your eyes as much as your spreadsheet.

Ultimately, the goal is to find those 3-5 spots per season where your assessment of the probable “warps” diverges meaningfully from the implied probability in the betting line. It’s not about picking every single over/under; that’s a fool’s errand. It’s about disciplined selectivity. For the upcoming season, I’m already looking at a team like the San Antonio Spurs. The line will likely be low, maybe 28.5 wins, based on their young core. But with a full year of Wembanyama’s development, the potential warp introduced by his defensive impact alone—which could transform them from a bottom-5 defense to a middle-of-the-pack unit faster than people think—makes the over intriguing to me. Conversely, a veteran-laden team coming off a deep playoff run might see an inflated total, not accounting for the cumulative fatigue and the higher injury probability that acts like a hidden trap on the track. The excitement lies in the dynamic interplay between the known outlines and the unexpected shifts. By building a model that is flexible, incorporating both hard data and a scout’s intuition for looming change, you position yourself not to predict every turn, but to successfully navigate the warps. That’s how you find value in the beautiful, unpredictable chaos of an 82-game NBA season.