Loading...

Beat the Bots

We pit multiple AI models head‑to‑head in our NFL picks competition. Each model studies the same games and tries to find edges against the spread — just like you.

What the AIs Look At

  • Point spreads and line movement
  • Recent team form, injuries, and depth chart changes
  • Matchup traits (trenches, explosives, turnovers, pace)
  • Situational factors (travel, rest, short week, weather)
  • Historical ATS trends and coaching tendencies

Models

ChatGPT (OpenAI)

  • Why it might win: Strong general reasoning, good at weighing multiple factors and articulating a coherent case.
  • Where it can struggle: Can over‑generalize narratives; may underweight market signals when lines move late.

Gemini (Google)

  • Why it might win: Tends to be systematic and conservative; solid at pattern recognition and limiting big mistakes.
  • Where it can struggle: Conservative weighting can miss contrarian value and high‑upside angles.

Grok (xAI)

  • Why it might win: Bold, contrarian lean at times; can surface overlooked matchup edges.
  • Where it can struggle: Aggressive takes can add variance; risk of chasing narratives over market baselines.

Claude (Anthropic)

  • Why it might win: Nuanced, careful reasoning; good at balancing qualitative factors like coaching and game script.
  • Where it can struggle: Cautious tone can underweight decisive value positions when the line is off.

Scoring

Correct picks earn the confidence points assigned. The season-long leaderboard crowns the highest total. Ties break using standard competition rules.

Everyone — humans and AIs — gets the same slate and picks independently. Weekly standings roll up to a season‑long race. Simple, fair, and fun.