Trading Concept
Correlation Breakdown
The tendency for assets that move independently in normal markets to become correlated during crises — eliminating diversification exactly when it is needed most.
During normal markets, Bitcoin and equities may have low correlation. Prediction market positions may move independently of crypto holdings. Card prices on TCGPlayer may have no relationship to forex rates. A user running bots across these markets may believe they are diversified — and during calm periods, they are.
During stress, correlations converge toward 1. Kristin Forbes and Roberto Rigobon documented this rigorously in their 2002 paper in the Journal of Finance. The mechanism is straightforward: when fear rises, investors sell everything liquid to raise cash or reduce exposure. The selling pressure hits all asset classes simultaneously. The diversification that protected in calm periods evaporates when protection is needed most.
This matters for anyone running multiple bots across assets they believe are uncorrelated. During a market stress event, all positions may move against the user simultaneously. The total portfolio drawdown can exceed what any single-strategy analysis would predict because the strategies are no longer independent. The tolerance band on each individual bot does not account for the aggregate exposure across all bots. Portfolio-level risk management — how much total capital is deployed across all running bots — is the user's responsibility.