We will be undergoing planned maintenance on January 16th, 2026 at 1:00pm UTC. Please make sure to save your work.

Kalshi Anomalies - Maduro's capture

Several traders made hundreds of thousands of dollars betting big BEFORE the U.S. captured Maduro. Coincidence? My statistical analysis says no. See the evidence inside.

Comment

Inspiration

At 9:21 UTC on January 3, 2026, President Trump announced the capture of Nicolás Maduro. Within minutes, Kalshi prediction markets exploded. But when I looked at the data, something didn't add up: massive YES bets had already been placed hours before the operation even started.

Was this luck, or did someone know what was coming?

What it does

My analysis detects statistically significant trading anomalies in Kalshi's "Maduro Out?" prediction markets. I identified:

  • $764,610 in theoretical profits for YES traders on January 3rd alone
  • A 10.6x spike in trading volume hours before the operation began
  • A suspicious cluster of large trades between 03:45-06:00 UTC—when intelligence operatives would have known the mission was underway
  • Trade size distributions that are statistically different from baseline (p < 0.001)

How I built it

  1. Data extraction: Downloaded 17,162 trades via Kalshi's public API
  2. Anomaly detection: Built 4 detection algorithms—large bets (Z-score), volume spikes, whale clusters, and timing anomalies
  3. Statistical validation: Applied Kolmogorov-Smirnov and chi-squared tests to confirm significance
  4. Visualization: Created interactive Plotly charts to tell the story visually

All code runs in a Hex notebook for reproducibility and collaboration.

Challenges I ran into

  • No trader IDs: Kalshi's API doesn't expose user identifiers, so we can only flag what happened, not who did it
  • Defining "suspicious": I had to calibrate thresholds (3σ, 15% price moves, 2-min windows) to reduce false positives while catching real anomalies
  • Correlation ≠ causation: Did traders have inside info, or did news leak through Venezuelan social media before Trump's announcement?

Accomplishments that I'm proud of

  • Built a reusable anomaly detection framework that can be applied to any Kalshi market
  • Found statistically significant evidence of unusual trading—not just vibes, but p-values < 0.001

What I learned

  • Prediction markets are powerful, but information asymmetry can undermine their legitimacy
  • Public trade data alone can reveal suspicious patterns—imagine what regulators with trader IDs could find
  • The window between a covert operation starting and public knowledge is a goldmine for informed traders

What's next for Kalshi Anomalies - Maduro's capture

  1. Expand to other markets: Apply the same framework to other high-impact events (elections, Fed decisions, geopolitical crises)
  2. Real-time monitoring: Build a live dashboard that flags anomalies as they happen, not after the fact it does

Built With

Share this project:

Updates