Data Pipeline

Key parameters that inform margining, hedging, and allowable exposure.

The engine continuously evaluates the underlying prediction market to determine safe leverage and notional. Inputs fall into four categories: liquidity, pricing, expiry, and event structure.

Liquidity Inputs

• Orderbook depth

  Executable depth within the configured band for both YES and NO.

• Spread

  Best bid/ask width; used to estimate hedging cost.

• Flow velocity

  Frequency of updates and trades; low activity increases execution uncertainty.

• Stability

  Detection of stale books, frozen quotes, or erratic update patterns.

Pricing Inputs

• Probability level

  Determines how much leverage is theoretically safe (lower p → more room).

• Recent volatility

  Measures short-term price variance to model drift and jump risk.

Expiry Inputs

• Time to expiry

  Less time → greater jump risk → stricter leverage.

• Expiry structure: fixed vs uncertain resolution date

  • Fixed-date markets allow predictable risk decay.

Event-Structure Inputs

• Binary vs gradual information release

  • Some markets reprice in small increments (polling shifts, economic releases). Others exhibit binary, instantaneous repricing (court rulings, announcements, results). The engine reduces allowable leverage for markets with historically fast, discontinuous moves.