Algorithmic Sports Betting

A professional sports bettor explains how he built and runs an automated betting operation that functions similar to an algorithmic hedge fund.

Core Business Model

• Fully automated betting system with zero human intervention
• Uses two parallel models running simultaneously
• Focuses on NBA betting with automated programs
• Measures success by ROI per bet rather than annual returns
• Performance metrics:
  • Bad year: 4-4.5% ROI per bet
  • Good year: 8-9% ROI per bet

Team Structure & Evolution

• Started as a small operation:
  • Main decision maker
  • One quant ("The Wiz") to build models
  • Two people handling bet placement and lineups
• Current structure:
  • ~250 people in organization (excluding football team)
  • Runs automatically with minimal oversight
  • Employees have significant profit sharing

Betting Operations

• Uses "beards" to place bets
  • Never places bets directly
  • Employs various types as fronts:
    • Successful business people
    • Degenerate gamblers
    • Hollywood people
    • Professional athletes (briefly used Floyd Mayweather)
• Avoids casino betting
  • Casinos are "for losers who can't win"
  • Uses street bookmakers and credit betting
• Maintains anonymity
  • Nobody knew identity of winning bettor
  • Uses multiple layers to hide betting activity

Historical Context

• References Billy Walters as pioneer
  • First to use computer programs
  • Used early Westinghouse machines
  • Had significant edge over manual odds makers
• Modern evolution
  • Fully automated systems
  • No human override on model decisions
  • Focuses on volume and consistent ROI