Reproducing a sudden, extreme, short-lived price collapse and recovery — the kind of event that has repeatedly exposed weaknesses in automated trading systems. Simulating one lets a team see how their algo actually behaves under panic conditions instead of hoping it never finds out live.