Shooting Luck Regression (3PT%)

Overview

This pillar identifies NBA players on unsustainably hot or cold three-point shooting streaks. It uses shot quality data to determine when a player's performance is driven by luck and is statistically likely to regress to their career average.

What It Does

The pillar calculates a player's expected 3-point percentage (x3P%) based on objective factors like defender distance, shot location, and shot type. It then compares this expected value to the player's actual shooting percentage over a recent stretch of games. A large variance between actual and expected percentages signals a high probability of future regression.

Why It Matters

Public markets often overreact to recent performance, creating value in betting against outliers. This pillar provides a statistical foundation to fade players on temporary hot streaks or back players in a slump, profiting from the inevitable return to the mean.

How It Works

First, the model ingests shot-by-shot data for every NBA player, including defender proximity. It then calculates a season-long x3P% for each player based on their typical shot quality. This baseline is compared against their actual 3P% over the last 5-10 games to create a 'Luck Variance' score, highlighting top regression candidates.

Methodology

The core formula is Luck Variance = (Actual 3P% over last 10 games) - (Season-long Expected 3P%). Expected 3P% is derived from a model using NBA optical tracking data, primarily factoring in defender distance classifications (e.g., tight, open, wide open) and shot location on the court. A variance greater than +12% or less than -12% is a strong signal.

Edge & Advantage

This provides a quantitative edge by systematically identifying and exploiting the market's recency bias, allowing for profitable bets against players whose recent performance is driven by luck, not skill.

Key Indicators

• 3P% vs x3P% Variance

  high

  The difference between a player's actual 3-point percentage and their expected percentage based on shot quality.

• Shot Quality Score

  high

  A metric evaluating the average difficulty of a player's three-point attempts, based on defender distance and location.

• Rolling 10-Game 3P%

  medium

  The player's 3-point shooting percentage over their most recent 10 games, used to identify current streaks.

Data Sources

• NBA Stats API

  Official source for play-by-play and player tracking data, including shot charts and defender proximity.

• Second Spectrum

  Provider of advanced player tracking data used by NBA teams, the gold standard for shot quality analysis.

• Cleaning the Glass

  A public-facing analytics site that provides filtered stats and expected values for player performance.

Example Questions This Pillar Answers

• → Will Stephen Curry make over or under 4.5 three-pointers in his next game?
• → Will a player currently shooting 55% from three over their last 5 games shoot under 40% in their next game?
• → Will a team that has been hot from three see its overall point total go under the projected line?

Tags