The selected passage discusses limit orders, market orders and repeated short-term price movements before a sustained change. The author treats transaction behavior as a source of nonrandom properties to study.
Separate the quote from the transaction
Our simple example illustrates bid-ask bounce. Trades alternating between a bid and an ask can create reversals in last-traded prices even when the midpoint is unchanged. That is a measurement issue, not a complete explanation of all price moves.
Worked example
Suppose a fictional quote remains bid $99 and ask $101. A sale at $99 followed by a purchase at $101 changes the last price by $2 while the midpoint stays $100.
Case connection
Investigators looked beneath the price chart at market depth. Transaction mechanics belong in the explanation.
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
Last-traded prices are observations of transactions, not a complete picture of available liquidity. Compare them with quotes, depth and the timing of actual orders. In a simple bid–ask example, alternating trades can create apparent reversals without a changing midpoint. The crash investigation encourages looking for such measurement distinctions before choosing a narrative explanation.
Try it
Construct four alternating transactions at an unchanged bid and ask. Compare their returns with midpoint returns.