GameStop: The Short Squeeze That Rewrote the Rules
In January 2021 a dying mall retailer became the center of world finance for two weeks, and a hedge fund giant never recovered. Part of our Looking Back series on 2020 to 2026, written from 2026.
The Setup
At the beginning of 2021 GameStop was a mall video game retailer in structural decline and Wall Street had bet on that decline as much as a bet can be made. short sale means borrowing shares selling them and hoping to buy them back cheaper later and GameStop short interestMore shares were sold short than existed freely which is possible because borrowed shares can be relent and sold short again and created the most explosive setup in the markets because every short seller is a guaranteed future buyer
On the other side were the retail traders on Reddit's WallStreetBets forum most visible among them Keith Gill posting as Roaring Kitty who had spent months arguing that stocks were undervalued and shorting was mathematically unwise. They were stuck at home full of stimulus checks and no trading commissions on apps made to look like games. The firewood had been piling up for a year as described in previous posts in this series. GameStop was thematch
A Worked Example: Why 140 Percent Short Interest Is a Trap
The phrase every short seller is a guaranteed future buyer is the crux of the entire episode and it becomes obvious the moment you attribute the stock count to it
Start with the float. GameStop's freely tradable share count stood in the region of 50 million shares after excluding insiders and long-term index holders who never sell. Short interest near 140 percent of that means roughly 70 million shares had been sold short
Now count how many shares should ultimately be purchased. Each of those 70 million shares is a purchase obligation. It is not a possibility it is an obligation because a short position is closed by buying the share. Thus there were 70 million shares of forced future demand compared to a set of around 50 million shares that could actually be delivered
Then divide by liquidity. GameStop traded between 5 and 10 million shares on a typical day before this began. Let's take 5 million as a base case. Seventy million shares of coverage at five million shares per day is fourteen trading days of buying assuming the short sellers are the only buyers in the market and no one else is competing with them for the shares
| Quantity | Approximate figure |
|---|---|
| Freely negotiable float | around 50 million shares |
| Shares sold short by 140 percent | around 70 million shares |
| Forced future purchase | around 70 million shares |
| Normal daily volume | around 5 million shares |
| Days of normal volume needed to cover | about 14 |
Read the last row and January's price action will no longer seem like a mystery. Fourteen days of buying had to occur at some point from participants who couldn't choose the moment once losses forced them to act to an action where other buyers arrived simultaneously and existing holders refused to sell
The position did not become dangerous when Reddit realized it. It was structurally dangerous the day it was put up and the arithmetic was published daily with data that any subscriber could see
One more calculation explains the recklessness. A short position has an asymmetric profit that goes in the opposite direction. If you short a stock at $17 the best possible outcome is for it to go to zero earning $17 per share. The worst outcome has no limit. At $483 a short seller who entered at $17 lost $466 per share approximately twenty-seven times the maximum possible profit from the same trade
These are illustrative figures using round numbers for float and volume. The structure holds regardless: unlimited downside limited upside forced buying greater than marketable supply and a fourteen-day exit queue
Two Weeks in January
The stock entered January 2021 around $17. As buying pressure increased mechanics took control. Shorts forced to cut losses had to buy shares driving up the price forcing the next short to buy the classic short squeezeAbove ran a gamma compressionOn January 28 the stock traded $483 intraday about twenty-five times higher in a month and GameStop was briefly among the most traded securities in the world
Melvin Capital the largest publicly shorted hedge fund lost 53 percent of its money in January alone and received an emergency injection of $2.75 billion from Citadel and Point72 amid the squeeze. The fund never regained its footing and closed in May 2022. The forum had set out to burn a short seller and actually did so which has never happened on this scale in the age of social media
The Day the Buy Button Vanished
On the morning of January 28 Robinhood and several other brokers restricted the buying not the selling of GameStop and other meme stocks. It seemed to users that the establishment was bailing out the hedge funds by pulling the plug on the crowd. The boring truth that was revealed later was the plumbing. Stock trades take time to settle and the industry clearinghouse requires guarantees from brokers against unsettled trades escalations to volatility and concentration. GameStop's numbers destroyed thoseformulas the clearinghouse demanded billions in deposits that Robinhood didn't have overnight and restricting purchases was the broker's way of reducing its exposure to survive. Robinhood raised more than $3 billion from investors in a few days just to reopen the button
The squeeze taught a generation the true market chart. Between the touch of a phone and the exchange of stocks lie brokers market makers clearinghouses and collateral formulas and in a crisis the pipelines trump them all both retail trading and hedge funds
Case Study: Volkswagen, 2008
The most useful thing to know about January 2021 is that it had already happened on a larger scale to more sophisticated participants thirteen years earlier
Throughout 2008 hedge funds had built a large short position in Volkswagen common stock on a thesis that seemed sensible: VW was trading at a high valuation relative to Porsche which was chasing it and a global financial crisis was crushing demand for automobiles
On October 26 2008 Porsche revealed that through a combination of stock and cash-settled options it controlled approximately 74 percent of Volkswagen. The German state of Lower Saxony owned about 20 percent and was not selling. If subtracted the free float available in the market was something like 6 percent compared to a short position of approximately 12 percent
That is the identical structure to the previous table in a much more extreme form. You had to buy twice as many shares as there were to buy
What followed lasted two days. Volkswagen shares went from around 200 euros to more than 1,000 euros intraday and the company briefly became the most valuable in the world by market capitalization above Exxon more by a mathematical impossibility than by an insight into cars. Hedge funds lost billions of euros. Porsche finally released some shares to relieve the pressure
Three things carry over directly to 2021. The participants were professionals and not retailers so this is not a story about unsophisticated crowds. The trigger was a revelation about ownership rather than any change in the business. And the losses were not due to being wrong about the company since VW was really well valued but to being wrong about whether the shares would be available at a price when the position had to be closed
Volkswagen should have made GameStop unthinkable. It was thirteen years old it was widely documented and taught in risky courses. The position was taken anyway which is the part worth sitting down with
What Actually Changed
Congressional hearings came and went but the lasting changes were structural. Liquidation the delay caused by the collateral crisis was shortened from two days to one by 2024. Payment for order flow the practice by which brokers sell retail orders to market makers received lasting regulatory scrutiny. Hedge funds learned to treat crowded short positions and social sentiment as inputs to position sizing and never again have short interests been allowedabove the float to pile on so blatantly to a single name. And retail traders for better or worse learned that they could act as a coordinated force a fact that every subsequent meme cycle including GameStop's own aftermath has demonstrated
Where the Story Gets Told Wrong
GameStop generated more myths per event unit than anything else in modern markets and four of those myths are worth correcting
Retail trade did not make the majority of purchases. The narrative is that of small investors against Wall Street. The volumes involved were far greater than retail order flow alone could produce and institutional participants including hedge funds that spotted the setup did much of the buying. Some of the money made in January 2021 came from exactly the type of company the story portrays as the villain
The buy button restriction was not a conspiracy and the conflict is still real. The clearinghouse collateral explanation is correct and well documented. It is also true that Robinhood's revenue depended on the flow of sell orders to a company that had just invested in Melvin and that a business model with that structure invites exactly this suspicion. Both things are true and the honest position is that the specific allegation was incorrect while the underlying conflict deserved the scrutiny it received
Short interest above 100 percent is not evidence of fraud. It is mechanically derived from stock lending. A borrowed stock can be sold to a new owner who lends it back so the same stock supports multiple short positions. That makes the position dangerous as the above arithmetic shows. That does not mean it is false or illegal and many unsuspecting commentators in 2021 were wrong
Most of the participants lost money. The stock traded at $483 intraday on January 28 and very few people sold anywhere near that figure. Studies of retail brokerage activity during the meme stock period generally find that most participants who bought during the frenzy were underwater afterwards. A famous victory by a crowd is consistent with most members of that crowd losing
How I Actually Read a Crowded Trade
The transferable skill of this episode isn't about meme stocks. It's about recognizing when the exit from a position is more dangerous than its thesis
The first thing I look at is days to cover which is short interest divided by average daily volume exactly the calculation in the example above. It is published on most data providers and converts a percentage into a number of days spent trying to exit. Anything in the double digits means the exit is the risk regardless of whether the thesis is correct
Second I look at who owns the shares that are not shorted. A float dominated by holders who will not sell at any price whether a government a founder or an ideologically committed retail base effectively reduces the float well below the reported number. This is what Lower Saxony did in 2008 and what the forum did in 2021
Third I try to separate the question of whether a company is bad from the question of whether the trade is good. GameStop was a failing mall retailer and Volkswagen was expensive. Both theses were defensible and both trades were catastrophic because being right in business doesn't help if you can't hold your own along the way
Fourth I treat asymmetrical payoffs with real suspicion. Limited wins and unlimited losses are a form that produces many small wins and occasionally full wins which flatters a track record until it doesn't
None of this is advice and I have never shorted anything
The Bottom Line
GameStop consisted of three stacked stories: a legitimately reckless short position that ran up against the mechanical physics of squeezes a social media crowd that discovered it could move markets and a compensation system that no one understood until it failed
The recklessness was more arithmetic than opinion. About 70 million shares were sold short against a float of about 50 million in a stock that trades approximately 5 million shares per day which meant that fourteen days of forced buying had to occur through a bid that did not exist with a short seller's maximum profit of $17 per share against a loss that reached $466
Volkswagen had already proven it all in October 2008 when Porsche's disclosure left a 6 percent free float versus a 12 percent short position and the shares went from around 200 euros to more than 1,000 in two days. In the end the shares themselves mattered less than the episode exposed. The markets are a machine with pipes incentives and bottlenecks and the two weeks in January 2021 in which a mall retailercommercial surpassed Apple in commerce remain the best free lesson ever offered on how that machine really works