Shooting Luck Regression

Overview

This pillar analyzes defensive shooting statistics to find teams that are either overly lucky or unlucky. It identifies teams whose opponents are shooting at unsustainable rates, providing an edge before the market corrects.

What It Does

It compares a team's opponent three-point and free-throw percentages against national averages and shot quality data. The model flags significant deviations, suggesting that a team's defensive performance is inflated or deflated by variance. This isolates teams primed for regression to the mean in upcoming games.

Why It Matters

Shooting luck can temporarily mask a team's true defensive strength, creating inaccurate market lines. This pillar provides a data-driven signal to bet against teams on an unsustainable lucky streak or back teams suffering from bad luck before performance normalizes.

How It Works

The pillar first ingests game-by-game defensive shooting data for all NCAA Division I teams. It then calculates the z-score for each team's opponent 3P% and FT% relative to the national average. Teams with scores beyond a certain threshold are flagged as regression candidates, with a higher score indicating a more significant deviation from the norm.

Methodology

A 'Regression Score' is calculated using a weighted average of z-scores for opponent 3-point percentage (Opp 3P%) and opponent free throw percentage (Opp FT%) over a rolling 15-game window. The formula is: Score = (Z_Opp3P% * 0.7) + (Z_OppFT% * 0.3). Teams with a score > +1.75 are flagged as 'Unlucky' (positive regression expected), while those < -1.75 are 'Lucky' (negative regression expected).

Edge & Advantage

This pillar quantifies luck, a factor most bettors assess subjectively, providing a clear signal to fade teams whose defensive records are artificially inflated.

Key Indicators

• Opponent 3P% vs. National Average

  high

  Measures how much better or worse opponents shoot from three against a team compared to the D1 average.

• Defensive Shot Quality (eFG%)

  high

  An estimate of the expected field goal percentage a defense should allow based on the location of shots taken.

• Opponent Free Throw Percentage

  medium

  While less predictive, significant long-term deviations from the mean in opponent FT% can indicate luck.

Data Sources

• KenPom

  Provides advanced analytics for NCAA basketball, including opponent shooting percentages and a proprietary 'Luck' rating.

• BartTorvik

  Offers free advanced statistical tools and queryable databases for college basketball analytics.

• ShotQuality

  Specializes in providing expected vs. actual make percentages based on shot location and type.

Example Questions This Pillar Answers

• → Will the total points in the Kansas vs. Baylor game go Over/Under 148.5?
• → Will Duke cover the -6.5 point spread against Virginia?
• → Will Houston hold their next opponent to under 30% from three-point range?

Tags