Statistical Arbitrage Strategy¶

The Statistical Arbitrage (Stat Arb) strategy exploits price divergences between correlated prediction markets.

How It Works¶

Identify correlated markets - Find markets that historically move together
Monitor spread - Track the price difference between correlated pairs
Trade divergence - When spread exceeds threshold, bet on convergence
Exit on convergence - Close positions when prices realign

Example¶

Market A: "Will candidate X win state Y?" - YES at $0.60
Market B: "Will candidate X win state Z?" - YES at $0.45

Historical correlation: 0.85
Current spread: 15% (above 4% threshold)

Strategy:
- Buy YES on Market B (underpriced)
- Optionally sell YES on Market A (overpriced)
- Wait for convergence

Configuration¶

# Minimum spread to trigger a trade
STAT_ARB_SPREAD_THRESHOLD=0.04

# Hours of price history for correlation
STAT_ARB_LOOKBACK_HOURS=24

# Minimum correlation coefficient
STAT_ARB_MIN_CORRELATION=0.7

Risk Level¶

Medium - Correlation can break down, especially during major events.

Risks: - Correlation breakdown - Markets may diverge further before converging - Liquidity differences between markets - Event risk affecting one market but not the other

CLI Commands¶

# View computed correlations
polybot statarb correlations --min-corr 0.7

# Manually compute correlations
polybot statarb compute --hours 48

# Show current opportunities
polybot statarb opportunities --spread 0.04

# Enable the strategy
polybot strategy enable stat_arb

Finding Correlated Markets¶

PolyBot automatically computes correlations between active markets. Good candidates:

Related political outcomes (same election, different states)
Sequential events (quarters, months)
Correlated economic indicators
Related sports outcomes

Best Practices¶

Verify correlation logic - Ensure markets should actually be correlated
Check liquidity - Both markets need sufficient volume
Monitor actively - Correlations can break during major news
Use position limits - Don't overexpose to correlation risk