REIGN — NBA Analytics

Quantifying NBA player impact across 80 years of basketball using era-specific composite models.

Research Paper

The full methodology is documented in our formal research paper:

REIGN: A Composite Metric for Quantifying NBA Player Impact Across Eras

Samir Kerkar — Courtside Analytics, March 2026

Key contributions:

• Four era-specific regression models (Pioneer, Legacy, Classic, Modern)
• 60/40 dampened blend for Modern era recomputation using scraped advanced stats
• Playoff opponent strength adjustment based on opposing team quality
• 29,969 player-seasons, 3,484 players, 1946–2025

All-Time Top 10 by Peak REIGN

#	Player	Year	Peak REIGN
1	LeBron James	2013	+27.70
2	Michael Jordan	1988	+26.59
3	Stephen Curry	2016	+26.59
4	Shai Gilgeous-Alexander	2025	+24.72
5	Nikola Jokic	2025	+24.40
6	Chris Paul	2009	+24.18
7	Kevin Garnett	2004	+24.13
8	James Harden	2019	+24.06
9	David Robinson	1994	+23.88
10	Shaquille O'Neal	2000	+23.81

How REIGN Works

REIGN uses different models for different eras because the available statistics differ fundamentally across NBA history:

• Pioneer (1946-62): TS%-dominant model, 11 features, no steals/blocks/BPM
• Legacy (1963-95): WS/48-dominant, 18 features, richest data
• Classic (1996-2012): WS/48 + VORP, 19 features including 3P%
• Modern (2013-25): WS/48-dominant after enrichment, 19 features

All scores are era-normalized via z-score within rolling 5-year windows, ensuring +20 REIGN means the same relative dominance whether it's 1988 or 2024.

Features

• Leaderboard — Sortable rankings with era filtering, 1yr/3yr/5yr peak views, clutch stats
• Player Profiles — Career arcs, skill radar, season heatmaps, award shelf for 3,484 players
• Head-to-Head Compare — Side-by-side stats, peak bars, trajectory overlay, radar, PNG export
• Era Explorer — Deep-dive into each era with evolution charts and cross-era tables
• Visualizations — League trends, OFF vs DEF scatter, age distribution, REIGN distribution
• RS/PO Toggle — Every page supports Regular Season and Playoff modes
• Opponent-Adjusted Playoffs — Scale playoff REIGN by opposing team quality
• URL-Shareable Comparisons — ?v=compare&p1=LeBron+James&p2=Michael+Jordan

Tech Stack

• React 19 + Vite 7
• Custom SVG charts + Recharts
• IBM Plex Sans + Source Serif 4 typography
• Service worker for data caching (stale-while-revalidate)
• Fuzzy player search (accent-tolerant, typo-friendly)
• html2canvas for PNG export (lazy-loaded)

Development

npm install
npm run dev

Build and Deploy

npm run build

Output goes to dist/. Works with Vercel, Netlify, or any static hosting.

Vercel (recommended): Push to GitHub, import in Vercel, deploy. No config needed.

Data

Split into parallel-loaded era files for fast initial load:

File	Size	Gzipped
seasons_pioneer.json	1.4 MB	0.2 MB
seasons_legacy.json	5.3 MB	1.1 MB
seasons_classic.json	6.5 MB	1.2 MB
seasons_modern.json	5.1 MB	0.8 MB

Additional: awards.json, stretches_rs3/rs5/po3/po5.json, career_avg_rs/po.json, career_clutch.json

Scripts

• scripts/regenerate_careers.py — Rebuild careers.json from seasons.json
• scripts/generate_paper.py — Generate the methodology PDF

License

Research and data for non-commercial use.