Postseason Prediction Models Update: Recalibrating October
October basketball is a beautiful, chaotic lie. We’re staring at sample sizes that barely qualify as a footnote, yet my models are already screaming for attention. If you ask me, the initial noise is exactly where the signal hides.
With a league-wide True Shooting percentage hovering at 57.2% through the first two weeks, the offensive efficiency explosion is real. It isn't just a fluke. When I look at the early Box Plus/Minus (BPM) leaders, the usual suspects are there, but the efficiency gaps are widening faster than I anticipated.
- Usage Rate vs. Efficiency: We’re seeing a 3.2% increase in high-usage players (30%+) maintaining a TS% above 60%. That’s historically unsustainable, yet here we are.
- Defensive Regression: Defensive Win Shares are down 12% across the board compared to this time last year. Teams are allergic to rotations, and the data proves it.
"The early season is a laboratory for variance. If you chase the outliers now, you’ll get burned. If you ignore them, you’re missing the shift in the baseline."
I’ve been recalibrating my projections daily. Shooting 48.9% from the field as a team is great, but when that’s tethered to a bottom-tier defensive rating, the win-loss record is destined to crater. Take a look at the current PER leaders; we have three players sitting north of 30.0. That’s a statistical anomaly that screams regression.
In my view, the teams currently sitting at .500 with a net rating near zero are the ones to watch. They aren't underperforming; they’re just waiting for their shooting splits to normalize. I’m not betting on the hot starts. I’m betting on the teams whose underlying metrics—the ones that don't show up in the box score—are quietly trending upward.
Numbers don't lie, but they do love to tease. We’re just getting started.
