Does Consistency
Win Championships?

Understanding Archetypes

Each team is classified into one of 16 archetypes based on four binary dimensions:

Examples:

• Elite Consistent: High offense + High defense + Consistent on both ends
• Elite Volatile: High offense + High defense + Inconsistent play
• Offensive Juggernaut: Great offense, poor defense, varies by consistency
• Defensive Anchor: Great defense, poor offense, varies by consistency

Core Metrics

Net rating is the difference between offensive and defensive efficiency, measured as points scored/allowed per 100 possessions.

Consistency Metrics

We test three different approaches to measuring consistency:

1. Rugosity - Arc-to-chord ratio measuring path volatility

A rugosity of 1.0 means perfectly consistent (straight line). Higher values indicate more volatile performance.

2. Standard Deviation - Classic volatility measure

Measures how much ratings vary from the average. Lower = more consistent.

3. Residual Standard Deviation - Volatility after removing trend

Captures "noise" around the trend. A team can be improving (positive trend) but still have high residual std if they're erratic.

Combined Metrics

To test whether consistency adds predictive power on top of raw performance:

Rewards teams that are both good AND consistent.

Penalizes inconsistency directly.

Trend Metrics

Trend scores capture whether a team is improving or declining over the analysis window.

We also calculate separate offensive and defensive trends to see if one side improving/declining matters more.

Archetypes

Each team is classified into one of 16 archetypes based on four binary dimensions relative to league medians:

• Offensive rating above/below median
• Defensive rating above/below median
• Offensive consistency above/below median
• Defensive consistency above/below median

Badge Analysis

We track which teams would have earned each badge in historical seasons and analyze their playoff performance:

• Consistently Good/Bad: Most consistent among top/bottom 10 net rating
• Best Offense/Defense: Highest offensive/defensive rating
• On the Rise/Decline: Highest positive/negative trend

Analysis Windows

We calculate metrics using the last 10, 20, 30, 40 games of the regular season, plus the full season. This tests whether late-season performance is more predictive than full-season averages.

Playoff Performance

Playoff success is measured by total playoff wins (0-16). Teams that missed the playoffs receive 0 wins. This creates a continuous scale where the champion has 16 wins.

Playoff Teams Only Analysis

We also analyze just playoff teams to answer: among playoff teams, does consistency predict deeper runs? This isolates the effect of consistency for teams that have already qualified.