The author examines the following-year behavior of poorly performing markets and asks whether price movements have regularities. The selected passage reports a historical comparison and a market-return table, not a universal instruction to buy every loser.
Reconstruct before extending
A proper follow-up would specify the included markets, ranking dates, return definitions and evaluation period. The sample can be sensitive to those choices. Our exercise focuses on making the claim reproducible rather than repeating its reported historical return as a promise.
Worked example
Suppose you rank five fictional markets at year-end and examine next-year returns. Choosing the universe after seeing the rebound would give the result an unfair advantage.
Case connection
One rapid recovery is a selected episode, not a general loser-rebound rule. Define the comparison set first.
The Flash Crash: price and liquidity
Source-grounded facts
E-mini futures and SPY fell about 5% within five minutes, then recovered over the next ten. Investigators examined order-book liquidity.
Context
On 6 May 2010, US equity and futures markets experienced a sharp, short-lived disruption. Investigators needed more than closing prices to reconstruct the episode.
Outcome
The price recovery did not erase the disruption. The investigation used trading records and liquidity data to distinguish the sequence of events from stories based only on a chart.
- Around 2:40 p.m., E-mini S&P 500 futures and SPY fell roughly 5% in five minutes, then recovered over the next ten.
- During the recovery, some individual stocks and ETFs traded at extremely low prices before rebounding.
- Staff examined full order books and found that futures buying depth had fallen dramatically; equity liquidity problems followed.
Case analysis
A dramatic rebound is easy to select after it happens. A fair rebound study starts with all qualifying declines and follows them under one horizon and rule. Include slow recoveries and cases that never recover within that horizon. This case can generate the hypothesis, but selecting it alone cannot supply the comparison needed to test that hypothesis.
Try it
Design a paper test of a loser-rebound hypothesis. Fix the universe and ranking period before reviewing subsequent outcomes.