The Flash Crashes (2010 & 2013): When Algorithms Went Rogue

By the 2010s, human floor traders screaming in the open outcry pits were largely a relic of the past. Wall Street had moved onto server racks located in data centers in New Jersey. Trading was now dominated by High-Frequency Trading (HFT) firms and complex algorithms executing thousands of orders per millisecond. This transition created unprecedented market efficiency, but it also introduced a terrifying new risk: the Flash Crash.

A flash crash is an event where the market drops a staggering amount in a matter of minutes—or even seconds—only to recover just as quickly, seemingly disconnected from any real fundamental economic news. Two of the most famous incidents perfectly illustrate the dangers of an entirely digitized market.

The Mother of All Flash Crashes: May 6, 2010

On the afternoon of May 6, 2010, the US stock markets were already jittery due to the unfolding European Debt crisis in Greece. The DOW was down modestly. Then, at 2:32 PM, the bottom completely fell out of the market in a way that defied the laws of financial physics.

In the span of roughly 36 minutes, the Dow Jones Industrial Average plunged an additional 600 points, hitting a loss of nearly 1,000 points (about 9%) for the day. Roughly $1 trillion in market value simply evaporated. Bizarre anomalies occurred across the market. The stock of consulting giant Accenture went from $40 to $0.01. Apple shares briefly traded at $100,000. It was pure electronic chaos.

By 3:07 PM, the market magically reversed. In another 20 minutes, the DOW recovered almost the entirety of the 600-point drop. It was as if the plunge never happened.

What Actually Happened?

It took regulators years to fully dissect the forensic data, but the crash was essentially caused by a lethal combination of a massive fundamental sell order and the predatory nature of HFT market makers.

• The Whale: A massive mutual fund (Waddell & Reed) decided to hedge its portfolio by selling a gigantic block ($4.1 billion) of E-Mini S&P 500 futures contracts. Their algorithm was programmed to simply execute the sale based on market volume, without regard for price or time.
• The Illusion of Liquidity: HFT market makers operate by constantly posting bids and asks, providing liquidity. However, when the Waddell & Reed algorithm started indiscriminately dumping billions in futures, the HFT algorithms smelled an anomaly.
• The HFT Retreat: HFT firms are not charities; their algorithms are programmed to instantly withdraw their bids if market volatility spikes to avoid getting run over. Within milliseconds, massive HFT liquidity entirely vanished from the order book. The Waddell & Reed algorithm kept blindly selling into an empty room, instantly driving the price into a black hole.

The "Hound of Hounslow"

Years later, the plot thickened. The US government arrested Navinder Singh Sarao, a British day trader operating out of his parents' house in West London. Regulators alleged that Sarao had significantly exacerbated the crash through an illegal practice called "Spoofing."

Sarao built a custom algorithm to constantly place massive, fake sell orders for E-Mini S&P 500 futures and cancel them a millisecond before they executed. This tricked the HFT algorithms into thinking there was massive institutional selling pressure, forcing them to lower their prices. Sarao simply scalped the difference. On May 6, 2010, his spoofing collided directly with the Waddell & Reed sell order, acting as the spark in a powder keg.

The 2013 Associated Press Tweet Crash

By 2013, the market had recovered and implemented new safeguards to prevent another 2010-style crash. However, Wall Street algorithms had evolved. They no longer just read moving averages and order flow; they were now using Natural Language Processing (NLP) to read the news—specifically, Twitter.

On April 23, 2013, the Syrian Electronic Army (a group of hackers) successfully compromised the official Twitter account of the Associated Press (@AP), a highly trusted global news wire.

At 1:07 PM, they tweeted: "Breaking: Two Explosions in the White House and Barack Obama is injured."

Human traders might have taken a few minutes to verify the news on CNN or Fox News. The algorithmic sentiment-analysis bots did not. Within milliseconds of the tweet hitting the API, automated systems across Wall Street recognized keywords ("Explosions", "White House", "Injured") and executed massive sell orders based on geopolitical terror logic.

In barely two minutes, the Dow dropped 143 points, wiping out roughly $136 billion in market cap. However, as soon as humans realized there were no explosions and the AP confirmed the hack, the algorithms were reversed. In exactly three minutes, the market fully recovered its losses.

The Lesson: Speed Kills

Both crashes definitively proved that modern algorithmic markets are fragile ecosystems. When humans remove themselves from the execution loop, the sheer speed at which computers can misinterpret data and trigger cascading feedback loops makes the financial system vulnerable to massive, unprecedented volatility spikes. Even with modern circuit breakers, the risk of a "Flash Crash" remains a permanent feature, not a bug, of digitized capital markets.