All events
Market Crashes
Flash Crash
For educational purposes only. Not financial advice. Higher returns come with higher risk. Never risk more than you can afford to lose.
Trading StatementREF: TS-2026-03
Trading StatementFor educational purposes only. Not financial advice. Higher returns come with higher risk. Never risk more than you can afford to lose.
By 2010, the structure of American financial markets had been fundamentally transformed by the rise of high-frequency trading. What had once been a market dominated by human specialists and floor traders on the New York Stock Exchange had evolved into a fragmented, electronic ecosystem where algorithms executed the majority of all equity trades. High- frequency trading firms, using co-located servers positioned mere feet from exchange matching engines, were capturing the spread on billions of shares per day, collectively accounting for an estimated 50-70% of all US equity trading volume. These firms argued that their activities improved market efficiency by narrowing bid-ask spreads and providing continuous liquidity, but critics warned that this liquidity was illusory and could vanish in times of stress. The debate over whether HFT was a net positive or negative for markets had been simmering for years and was about to be thrust into the spotlight.
The regulatory landscape had been shaped by Regulation NMS (National Market System), implemented in 2007, which required brokers to route orders to the exchange offering the best price. While intended to promote competition and reduce costs for investors, Reg NMS had the unintended consequence of fragmenting liquidity across dozens of exchanges and dark pools. A single stock could be traded on more than 40 different venues simultaneously, with high-frequency traders arbitraging price differences between them at microsecond speeds. This fragmentation meant that in normal conditions, markets appeared deep and liquid, but the depth could be misleading -- much of it consisted of HFT quotes that could be withdrawn in milliseconds. The old NYSE specialist system, for all its flaws, had carried an obligation to maintain orderly markets; the new electronic market makers had no such requirement.
The broader market environment on May 6, 2010 was already unsettled. The European sovereign debt crisis was escalating, with Greece on the verge of default and fears spreading about the solvency of Portugal, Ireland, Italy, and Spain. Earlier that week, violent protests had erupted in Athens, with three people killed when a bank was firebombed during an anti- austerity demonstration. The VIX volatility index had risen from 16 in mid-April to 27 by the morning of May 6, reflecting growing anxiety about European contagion and its potential impact on the global financial system. Markets had already declined approximately 4% from their April highs, and investor sentiment was fragile. The stage was set for a stress test of modern market structure, though no one realized it at the time.
The catalyst for the Flash Crash was a large sell order placed by Waddell and Reed Financial, a mutual fund company based in Overland Park, Kansas. Waddell and Reed's portfolio managers had decided to hedge their equity exposure by selling $4.1 billion worth of E-mini S&P 500 futures contracts -- approximately 75,000 contracts. This was an unusually large order, but not unprecedented in the E-mini market, which was one of the most liquid futures contracts in the world with average daily volume of approximately 1.6 million contracts. What made this particular order problematic was how it was executed, not its size. Similar orders had been placed and absorbed by the market many times before without incident.
Waddell and Reed used an automated execution algorithm that was programmed to sell at a fixed percentage of total market volume, without regard to price or time. The algorithm was designed to sell 9% of the trading volume each minute, completing the order over approximately 20 minutes. Critically, the algorithm had no price sensitivity -- it would continue selling at the same rate regardless of whether the market was stable or plunging. In a normal market environment, this approach would have been unremarkable. But on a day when the market was already under stress and liquidity was thinner than usual, the relentless selling pressure from this algorithm would interact with the behavior of high- frequency traders in a way that no one had anticipated. A more sophisticated algorithm with price awareness and volatility safeguards would likely have paused execution as conditions deteriorated.
Adding another layer of complexity to the setup was the activity of Navinder Singh Sarao, a self-taught trader operating from his parents' home in Hounslow, a suburb of west London. Sarao had been placing and rapidly canceling large orders in the E-mini S&P 500 futures market, a practice known as "spoofing" or "layering." By placing thousands of sell orders that he never intended to execute, Sarao created the illusion of heavy selling pressure, which influenced the behavior of other algorithms that interpreted the order book data as evidence of genuine supply. While Sarao's activities did not single-handedly cause the Flash Crash, they contributed to the deterioration of liquidity and the misleading signals that other market participants relied upon. His activities were estimated to represent a significant portion of the visible sell-side depth in the E-mini market during the critical minutes before the crash.
The timeline of the Flash Crash unfolded with breathtaking speed. At approximately 2:32 PM Eastern Time, Waddell and Reed's algorithm began executing its sell program in the E-mini S&P 500 futures market. The initial impact was modest -- the E-mini contract had already declined about 3% from its morning levels due to the European crisis fears. But as the algorithm continued selling at its fixed rate, the price of the E-mini began to fall more rapidly. High-frequency trading firms, which normally provide liquidity by continuously quoting both bids and offers, began detecting the unusual selling pressure and started reducing their quote sizes and widening their spreads. The early signs of trouble were subtle, visible only to those monitoring the order book depth in real time.
Between 2:41 and 2:44 PM, the situation deteriorated rapidly. As the E-mini futures price fell, index arbitrage algorithms automatically began selling stocks in the cash market to profit from the growing gap between futures and cash prices. This selling pressure spread from the futures market to individual stocks across all major exchanges. High-frequency market makers, observing the escalating volatility and the disappearance of normal correlations between related instruments, began withdrawing from the market entirely. One by one, the firms that normally provided the majority of visible liquidity pulled their quotes, leaving an increasingly thin order book that could not absorb the incoming sell orders. The withdrawal was rational from each firm's individual perspective -- they were protecting themselves from adverse selection -- but the collective effect was devastating.
At approximately 2:45 PM, the market entered what regulators later described as a "liquidity crisis." With HFT firms withdrawing and traditional market makers overwhelmed, sell orders cascaded through increasingly empty order books, finding buyers at ever-lower prices. In the E-mini market, the contract fell 5% in approximately four minutes -- an almost unprecedented rate of decline for the benchmark US equity futures contract. The selling created a "hot potato" effect, in which HFT firms that did remain active were rapidly buying and reselling contracts to each other in a frenzied loop, with each firm holding positions for mere seconds before passing them on, without providing any genuine absorption of the selling pressure. Volume surged but meaningful liquidity had vanished.
The most dramatic and disturbing manifestation of the crash occurred in individual stocks. As the cascade spread from futures to equities, some stocks experienced price dislocations that were almost impossible to comprehend. Shares of Accenture, a company worth over $30 billion, traded at one cent. Procter and Gamble, one of the largest and most stable companies in the world, fell from $60 to $39 in minutes. Conversely, some stocks were briefly quoted at absurd prices -- Apple and Sotheby's reportedly showed asks near $100,000 per share. These extreme prices resulted from the interaction of automated systems executing orders against stub quotes -- placeholder bids set at extremely low prices ($0.01) or extremely high prices ($99,999.99) by market makers who had effectively withdrawn from the market but had not technically removed all their quotes from the system. The stub quotes, which were never intended to be executed, became the only available liquidity in a market where all real buyers had disappeared.
Waddell and Reed Financial became the unexpected face of the crash, though the firm's role was more as a trigger than a cause. The Kansas City-based mutual fund company was a mid-sized, relatively conservative asset manager that had no intention of disrupting markets. Its portfolio managers made a reasonable decision to hedge equity exposure during a period of elevated uncertainty, but the execution algorithm they chose was poorly designed for the prevailing market conditions. The SEC and CFTC joint report on the Flash Crash, published in October 2010, identified Waddell and Reed's sell order as the initiating event, though it emphasized that the broader market structure vulnerabilities were the root cause of the extreme outcome. Waddell and Reed was not charged with any wrongdoing, and the firm later noted that it had executed similar orders many times before without any market impact.
Navinder Singh Sarao's role in the Flash Crash was not publicly identified until 2015, when the US Department of Justice charged him with spoofing and market manipulation. Sarao, who operated from a modest bedroom in his parents' home using a modified off-the-shelf trading platform, had allegedly been spoofing the E-mini S&P 500 market for years, generating tens of millions of dollars in profits. His arrest by British police at the request of US authorities was a sensational moment that raised uncomfortable questions about market surveillance. How had a single trader, operating from a London suburb with relatively unsophisticated technology, managed to manipulate one of the world's most important financial markets for years without detection? Sarao eventually pleaded guilty to spoofing and was sentenced to one year of home detention, with no prison time, partly because the judge acknowledged that he had cooperated with authorities and that his activities, while illegal, were not the primary cause of the Flash Crash.
The high-frequency trading industry came under intense scrutiny after the Flash Crash, with firms like Citadel Securities, Virtu Financial, and Jump Trading facing questions about their role in both providing and withdrawing liquidity. The industry's defense -- that HFT firms are not obligated market makers and should not be expected to provide liquidity during extreme events -- was technically correct but deeply unsatisfying to regulators and the public. If HFT firms accounted for the majority of displayed liquidity in normal times but disappeared during stress, then the market's apparent depth was fundamentally misleading. This realization prompted a broader debate about whether HFT firms that enjoy the benefits of market making (narrow spreads, high volume, co-location privileges) should also bear some of the obligations, including minimum quoting requirements during periods of volatility. The debate continues to this day and has shaped regulatory thinking across multiple jurisdictions.
The immediate market impact was astonishing in both its severity and its brevity. Between 2:42 and 2:47 PM, the Dow Jones Industrial Average plunged approximately 998.5 points, reaching an intraday low that represented a decline of roughly 9.2% from the previous close. Then, almost as quickly as it had fallen, the market reversed. The Chicago Mercantile Exchange's built-in safeguard, known as "Stop Logic Functionality," automatically paused trading in the E-mini contract for five seconds at 2:45:28 PM when the price declined too rapidly. This brief pause was enough to break the feedback loop. When trading resumed, buyers re-entered the market, HFT firms began quoting again, and prices rapidly recovered. By 3:07 PM, the Dow had recovered most of its losses, and it closed the day down 347 points, or approximately 3.2% -- a significant but unremarkable decline given the day's news about Europe. The five-second pause demonstrated the power of circuit breakers in disrupting cascading feedback loops.
The aftermath in individual stocks was particularly problematic. Exchanges were forced to cancel or "bust" more than 20,000 trades that had been executed at prices more than 60% away from their pre-crash levels. This meant that traders who had placed limit buy orders at extremely low prices -- hoping to scoop up stocks on the cheap during the crash -- had their trades reversed. Conversely, traders who had sold at the bottom through market orders or stop-loss orders saw those trades stand, locking in their losses. The inconsistent treatment of trades executed during the crash generated significant anger and legal disputes, and it highlighted the lack of clear, uniform rules for handling erroneous trades across the fragmented US market structure. The 60% threshold was seen as arbitrary, and many traders whose losses fell just inside the threshold felt they had been treated unfairly.
The Flash Crash caused approximately $1 trillion in market value to temporarily vanish before most of it was recovered within 36 minutes. But the psychological impact on investor confidence was lasting. Retail investors, many of whom had only recently returned to the stock market after the 2008 financial crisis, were shaken by the revelation that stock prices could collapse to a penny in seconds. The event contributed to a continued outflow from equity mutual funds and accelerated the shift toward passive index investing and ETFs, as investors sought to reduce their exposure to the kind of single-stock price dislocations that the Flash Crash had demonstrated were possible. Trust in market fairness, already damaged by the 2008 financial crisis, suffered another blow that would take years to repair.
The regulatory response to the Flash Crash was substantial and multi-layered. The SEC and CFTC jointly published their analysis of the event in October 2010, and the SEC moved quickly to implement new safeguards. The first response was the introduction of single-stock circuit breakers in June 2010, which paused trading in individual stocks that moved more than 10% within a five-minute period. These circuit breakers were later replaced in 2012 by the Limit Up-Limit Down (LULD) mechanism, a more sophisticated system that prevents trades from occurring outside of specified price bands that are recalculated every 15 seconds based on the stock's average price. The LULD mechanism represented a significant improvement over the previous system and has been credited with preventing several potential mini flash crashes in subsequent years. The system effectively eliminates the possibility of trades executing at absurd prices like one cent or $100,000.
The Flash Crash also catalyzed efforts to modernize market surveillance and crack down on manipulative trading practices. The SEC established the Market Information Data Analytics System (MIDAS) in 2013, which collects and analyzes billions of records of market activity daily to detect unusual patterns. The Dodd-Frank Wall Street Reform Act, signed into law in July 2010 (partly in response to the Flash Crash, though primarily driven by the 2008 financial crisis), included provisions that explicitly outlawed spoofing for the first time, defining it as "bidding or offering with the intent to cancel the bid or offer before execution." This provision was subsequently used to prosecute Navinder Sarao and numerous other traders engaged in similar practices. The Consolidated Audit Trail (CAT), a comprehensive system for tracking every order and trade across US markets, was also proposed in the aftermath of the Flash Crash, though its implementation has been plagued by delays and cost overruns.
Despite these regulatory improvements, the fundamental debate about market structure that the Flash Crash ignited remains unresolved. Critics of HFT, most prominently author Michael Lewis in his 2014 bestseller "Flash Boys," argued that the fragmented, high-speed market structure was fundamentally rigged against ordinary investors. Defenders countered that modern markets offer tighter spreads, lower commissions, and faster execution than at any point in history. The truth likely lies somewhere in between: electronic markets have genuinely reduced costs for most investors most of the time, but they have also introduced new categories of risk -- including the possibility of extreme, high-speed dislocations -- that did not exist in the old specialist-based system. The Flash Crash was a warning that the market's visible surface can mask hidden fragilities that only become apparent in the worst possible moment. Subsequent mini flash crashes in individual stocks and ETFs have reinforced this lesson.
The single most practical lesson from the Flash Crash is about order types. Traders who had placed market orders to sell during the crash received executions at catastrophic prices -- in some cases, a penny per share for stocks worth tens of dollars. Market orders provide certainty of execution but no certainty of price, and in a liquidity vacuum, the difference between the expected price and the actual execution can be enormous. Limit orders, by contrast, guarantee that you will not sell below (or buy above) a specified price, even if the trade may not execute at all if the limit is not reached. The Flash Crash demonstrated that during extreme events, a non-execution is vastly preferable to an execution at a price that destroys capital. Every trader should understand the difference between order types and default to limit orders for all but the most liquid, time-critical situations. This is perhaps the simplest and most actionable takeaway from the entire event.
The Flash Crash also teaches important lessons about the nature of liquidity in modern markets. The order book depth that is visible on a trading screen at any given moment is not a reliable indicator of the liquidity that will be available when you need to execute a large order during a period of stress. Much of the visible liquidity in today's markets is provided by HFT firms that can and will withdraw in milliseconds when conditions deteriorate. For traders managing significant positions, this means that exit strategies must account for the possibility that liquidity will evaporate precisely when it is most needed. Scaling out of positions gradually, using multiple venues, and avoiding concentration in less-liquid names are all practical measures that reduce exposure to liquidity risk. The apparent depth of the market is not its actual depth, and planning for the worst case is essential.
The role of automated execution algorithms in triggering the crash highlights the importance of building intelligence into trading systems. Waddell and Reed's algorithm sold at a fixed percentage of volume regardless of market conditions -- it had no awareness of the chaos it was helping to create. Modern execution algorithms should include multiple safeguards: price limits that pause execution if the market moves beyond expected ranges, volatility sensors that reduce participation during unusual conditions, and human override capabilities that allow a trader to intervene when the algorithm is behaving in unexpected ways. The principle extends beyond institutional algorithms to retail traders using automated stop-loss orders, which can trigger cascading executions at unfavorable prices during flash events. Any automated trading logic should include a "sanity check" that pauses execution when market conditions become abnormal.
Finally, the Flash Crash offers a broader philosophical lesson about the relationship between technology and risk in financial markets. Every technological advancement in market structure -- from the telegraph to electronic trading to high-frequency algorithms -- has simultaneously improved efficiency in normal conditions and introduced new categories of risk in extreme conditions. Traders operating in modern markets must accept that they are participating in a system whose complexity exceeds any individual's ability to fully comprehend. Humility about what can go wrong, combined with practical risk management measures like position sizing, diversification, and the disciplined use of limit orders, is the most reliable defense against the tail risks that the Flash Crash so dramatically revealed. The market can remain irrational far longer than you can remain solvent, and it can also become completely irrational far faster than you can possibly react. The 36 minutes of the Flash Crash compressed more drama and destruction than most traders experience in an entire career.