The 2010 Flash Crash
How a large automated futures sale met thinning liquidity and transmitted stress across linked equity markets.
Why this event still matters
On 6 May 2010, major US equity indexes fell with extraordinary speed and then recovered much of the move within minutes. The most striking prints were not confined to one instrument: stress passed among equity-index futures, exchange-traded funds, and individual shares. Some securities traded at prices that bore little relationship to their value immediately before or after the event. [1][2]
The official investigation did not identify a single broken machine or one malicious order as a complete explanation. It described an already fragile market, a very large automated futures sale, intermediaries unwilling to retain growing inventory, and linked venues transmitting the pressure. The case is therefore useful because it shows how ordinary market functions can interact to produce an extraordinary outcome. [2][3]
Event reconstruction
Automated selling begins
A large sell program starts executing against trading volume without regard to price or time.
Liquidity thins rapidly
High-frequency participants trade the same contracts back and forth while reducing net inventory.
Prices dislocate
Futures and individual securities move sharply as displayed depth becomes unreliable.
Markets recover
A trading pause and returning liquidity help prices move back toward prior levels.
A stressed market before the crash
The afternoon of 6 May 2010 did not begin in calm conditions. Concerns about European sovereign debt had already increased volatility, and broad US equity indexes were down before the most severe dislocation. That background matters because liquidity is adaptive: firms that normally quote both sides reduce size or widen prices when uncertainty and inventory risk rise. [1]
The joint SEC and CFTC investigation treated the event as a cross-market episode rather than a single erroneous trade. Equity-index futures, exchange-traded funds, and individual stocks were linked through hedging and arbitrage. Pressure in one venue could therefore cause trading in another venue even when the original order never touched that second market. [1]
Price, execution, and fair value are different
A print produced by a desperate order in an empty book can be a valid execution without being a stable estimate of value. This distinction explains how some securities traded at extreme prices and then rapidly returned. The matching engine records the best available counterparty at that instant; it does not certify that the price reflects a considered valuation by a broad set of investors. [1][2]
That matters for triggers tied to last trade, including stops and risk controls. A robust venue may use indexes, bands, or reference-price checks to reduce dependence on one anomalous trade, but every design involves tradeoffs. A slower reference can become stale, while a fast reference can transmit a temporary dislocation. Traders need to know which price governs margin and liquidation. [2][4]
The automated sell program
At approximately 14:32 Eastern Time, a large fundamental seller initiated a program to sell 75,000 E-mini S&P 500 futures contracts, valued at about $4.1 billion at the time. The algorithm targeted execution equal to 9% of trading volume and did not use price or time as a stopping condition. A similar program on an earlier occasion had been spread over several hours; on 6 May it completed in roughly 20 minutes. [1]
This does not mean one order explains every price. It means a large, mechanically paced flow entered an already defensive market. As volume accelerated, the algorithm also accelerated. A volume-based rule can therefore become procyclical: the market trades more because it is stressed, and the execution program responds to that higher volume by selling faster. [1]
Hot-potato trading and disappearing depth
High-frequency traders initially absorbed part of the futures flow, but their business model generally limited how much directional inventory they would retain. They rapidly sold contracts back into the market. The report described contracts being passed among intermediaries in a hot-potato pattern while aggregate buying interest from longer-horizon participants remained insufficient. [1]
Displayed volume can look active while genuine risk-bearing capacity is shrinking. A contract changing hands repeatedly is not equivalent to new buyers willing to warehouse the position. As inventories and volatility limits tightened, some liquidity providers paused, reduced size, or widened quotes. Market orders then reached successively worse prices with less quantity available at each level. [1][2]
How stress crossed into stocks and ETFs
Futures prices and cash-equity products normally remain connected through arbitrage. When the futures market fell, traders sold related ETFs and stock baskets or adjusted hedges. At the same time, fragmented equity-market safeguards and inconsistent liquidity conditions produced extreme executions in some individual securities. [1]
Some trades occurred at prices far from previous values, including near-zero prints and unusually high prints in individual instruments. These were not a uniform repricing of corporate value. They reflected orders encountering little or no executable interest at intermediate prices. The episode illustrates why a last trade is an event, not necessarily a robust estimate of fair value. [1][2]
The pause and rapid recovery
At 14:45:28, the E-mini futures market triggered a five-second Stop Logic Function after prices moved into a range with insufficient executable liquidity. That brief pause allowed indications of interest to accumulate. When trading resumed, prices stabilized and began recovering. [1]
The recovery was fast because the dislocation was partly about market mechanics and absent liquidity, not a comparable minute-by-minute change in fundamental value. Returning buyers, paused selling pressure, and restored cross-market links helped prices move back. A rapid rebound does not make the earlier executions unreal; traders whose orders filled during the gap still faced those prints unless trades were later cancelled under applicable rules. [1]
What a perpetual trader should carry forward
The durable lesson is not simply to avoid market orders. It is to understand that available liquidity depends on volatility, inventory, venue controls, and the behavior of linked markets. Position size should be chosen for stressed depth, not only the book visible during ordinary conditions. [1][2]
Perpetual venues add further feedback channels through mark prices, margin requirements, and automated liquidation. A sharp reference move can trigger forced orders that consume the same liquidity needed for discretionary exits. Stops, liquidation estimates, and displayed depth are planning inputs, not guaranteed execution boundaries. [1][2]
What changed after May 6
The event accelerated work on coordinated circuit breakers and later limit-up and limit-down protections for individual securities. These mechanisms are intended to create time for prices and liquidity to reform, but they are not a promise that every order will execute or that linked venues will pause in identical ways. [2][4]
Subsequent CFTC research also cautioned against reducing the episode to a slogan about high-frequency trading. Fast intermediaries initially supplied liquidity and later withdrew or traded aggressively as conditions changed. The relevant question is how inventory limits, execution algorithms, and market controls interact under stress. [3]
What to carry forward
The Flash Crash changed market safeguards, but it did not repeal liquidity risk. The enduring lesson is that the order book is a set of conditional intentions. During stress, the same participants who normally make execution easy may reduce size together. Traders should size positions for the depth that might remain after volatility rises, not the depth visible before it begins. [2][4]
Execution speed exceeded available depth
Linked venues transmitted price pressure
Displayed liquidity disappeared when it was needed most
Market depth is conditional, not guaranteed
A market order can behave very differently during stress
Cross-market connections can transmit liquidation pressure
Sources and primary documents
Claims above link to the numbered records used in this case file.
Expanded case file with a full introduction, deeper analysis, conclusions, and claim-level citations.
Report an issue with this case file