The Stock Market Always Wins

Source: The Stock Market Always Wins, Art of the Problem, 27:12, uploaded 2025-06-26, playlist index 232.

In 2007 Warren Buffett bet that a cheap index fund would beat a group of professional hedge funds over ten years. Art of the Problem uses the wager to ask how a strategy that does nothing can beat the people who spend their working lives trying to find an edge. The answer arrives through the market’s less visible workers. Traders who chase tiny discrepancies make prices agree with one another, move information between isolated markets, and leave the passive investor with a cheap claim on the result.

Buffett’s wager and the trader’s hidden job

Buffett’s chosen strategy is simple: buy every stock in proportion to its size, then leave the portfolio alone. The video says that the index fund won the 2007–2017 bet by a wide margin and has continued to beat most professional investors. That result creates a problem for the usual picture of finance. If the active traders are clever enough to price every company, why does the person who copies the whole market and pays almost nothing often come out ahead?

The video first frames traders as either parasites who extract money from ordinary investors or bees that perform a hidden service. The narrator begins from the first view, then changes his mind during the 2017 cryptocurrency bubble, when he accidentally became a small market maker.

The empty sell side

The narrator had followed Bitcoin from its early period as a hobby for geeks and libertarians. When the price passed $1,000 and then$2,000 in 2017, he joined several small exchanges. One exchange displayed Bitcoin at $1,850 whilst other exchanges showed roughly$2,000. He placed a $100 market order, which should have bought at whatever price the exchange could offer. The order stayed open because nobody had recently traded there.

The order book explained why. It is a set of price levels, with buy orders gathered on one side and sell orders on the other. A limit order waits at a stated price. A market order consumes the cheapest available limit orders and moves through the book when the first level cannot fill the whole amount. The fuller the levels, the more depth the market has and the less a large order moves the price.

This exchange had no sell orders at all. The narrator transferred some Bitcoin from his wallet and posted a small amount for $2,500, even though Bitcoin traded at$2,000 elsewhere. It sold at once. He raised the price, and the market kept buying. He then posted one Bitcoin for $10,000, partly as a joke, and it sold immediately. The exchange had many market-buy orders waiting for any available supply, so his limit orders became the whole sell side. He had cornered the market by accident.

When he ran out of Bitcoin, he bought more on a major exchange and waited through a fifteen-minute transfer. By the time the coins arrived, other sellers had found the same opportunity and filled the book with reasonable prices. Each time he posted a better sell order, an automated market maker appeared almost at once and undercut him by one cent. His lost opportunity was the ordinary buyer’s gain. Competition had brought the price back towards a level at which a market order could execute without much slippage.

The video connects this small crypto episode with the block-order problem of the stock market in the 1980s. A pension fund trying to buy $50 million of IBM stock could move the price by 8% because the book lacked enough depth. Market makers reduced the effect by keeping orders near the current price. The video presents the reported fall from several percentage points of slippage to a fraction of a percent as an example of the service that traders provide. It gives no independent source for those figures.

Arbitrage as market information

As more crypto exchanges appeared, the narrator found larger price differences. He asked a programmer friend to help build a bot that could search across exchanges and currencies. The video describes the 2017 market as a place where roughly $5 billion changed hands each day whilst large professional trading firms had yet to arrive. In moments of panic, the same coin could differ by 20% between exchanges.

The bot searched for circular routes. It could start with dollars, buy Bitcoin, exchange Bitcoin for Ethereum, then turn Ethereum back into dollars. If the exchange rates multiplied to a number above one, the loop contained a profit. On its first day, the bot turned $200 into an$8.50 gain in a second. The opportunity disappeared as soon as the trade executed because the trade itself brought the prices closer together.

The easy money lasted briefly. Other traders found the larger gaps, faster firms arrived with deeper pockets, and the remaining opportunities shrank to tiny margins. That competition also connected the exchanges. When news changed Bitcoin’s value in one place, arbitrage traders carried the new price through the rest of the network. A small crypto market became a model for a larger financial system in which traders constantly compare prices in different places and react to new information.

The video describes arbitrage through geometry. If Bitcoin has the same price on exchanges A and B, the pair sits on a diagonal line where the two prices agree. The line acts as a no-profit surface. A price difference moves the market point away from it, creating an opportunity to buy on one exchange and sell on the other. Each arbitrage trade pushes the point back towards the surface. With more currencies and exchanges, the surface gains more dimensions, although the relation stays the same.

Future arbitrage uses a different kind of surface. Traders estimate where prices are likely to be later, so the boundary becomes fuzzy because the future carries uncertainty. The period can be decades, days, or fractions of a second. A long-term investor such as Buffett and a high-frequency algorithm then occupy different parts of the same problem. Each tries to recognise a price that seems out of line with a likely future state.

Benjamin Graham, Coca-Cola, and the short-term trader

Buffett credits Benjamin Graham, whose 1949 book The Intelligent Investor teaches investors to research a company’s earnings, growth, and future profits. If the likely future value exceeds the current price, the investor buys and waits. Graham’s difficulty is the daily price itself. He personifies the market as Mr. Market, a visitor who arrives each day with a new offer and lets fear or enthusiasm carry him away from the company’s value.

The video uses Buffett’s Coca-Cola investment as an example of this long-term prediction. In the 1980s, many investors saw a mature American soda company. Buffett saw consumption rising in markets that had room to grow. The video gives roughly 300 Cokes per American each year, three in China, and fewer than one in India, with those latter markets beginning to climb as they opened. Buffett also trusted Coca-Cola’s plan to make the drink available around the world. The video says that by 1994 he had put 35% of the fund, or $1.3 billion, into the company. By 1998, it says, the investment had grown to$20 billion as Coke sales expanded globally.

The bet depended on shorter-term traders. They inject new information into prices every second, even though Graham tells Buffett to ignore their daily offers. Without their activity, the market would take longer to recognise the changes that make a long-term investment valuable. Paul Tudor Jones becomes the video’s example of this work. The footage shows him reacting to an OPEC production-cut agreement, weighing whether thirteen countries can keep a promise, and trying to place orders before the market has finished absorbing the news.

The same role appears in a prediction market around the 2025 papal election. The video identifies a trader as Domer and says that he noticed a bias against an American pope. When the white smoke appeared, the market’s two favourites rose sharply, whilst the trader had found Robert Prevost at odds of 200 to 1. Prevost then emerged as Pope Leo XIV. The source presents the trade as payment for research in a market short of information. The trader did not create the event. He made a neglected possibility visible to other participants.

Statistical arbitrage and the geometry of relationships

The next step replaces exact price equality with a relationship that usually holds. The video credits Gary Bamberger at Morgan Stanley with pioneering statistical arbitrage in the 1980s. Coke and Pepsi serve as the example. Their prices tend to move together because their businesses meet a similar demand, so their usual relation forms a fuzzy corridor rather than a sharp line.

When Coca-Cola launched New Coke in 1985 and consumers rejected it, Coke fell whilst Pepsi benefited. A statistical-arbitrage strategy could buy the damaged Coke shares and sell Pepsi, betting that the pair would return towards its usual relation. When Coke later reports strong sales and rises by 10%, the same strategy asks why Pepsi and other related companies should remain unchanged. The traders spread information from one stock to the companies that ought to respond to it.

D. E. Shaw scaled this idea across thousands of stocks and traded many of them together. The relationships become a map of the economy. Bank stocks form one broad group. A change at Coca-Cola can reach Pepsi, other drinks companies, restaurants, and sugar suppliers. Algorithms model those connections as curved regions in price space, then trade when a group moves away from a pattern that the model expects to hold.

Jim Simons gives the argument its most mathematical form. His earlier work studied geometric distortions and measured when points moved away from their expected place on a surface. Renaissance Technologies treated markets as another geometric system. Prices followed shapes imposed by the real world’s constraints, and a deviation created a trade that could profit when prices returned towards the expected region.

Simons’s firm avoided the usual finance pedigree. He says that a physicist can learn finance, whilst a finance specialist cannot easily learn physics. The models treated prices as time series and searched across hundreds or thousands of them for relationships. They later absorbed prices, news, weather, economic data, satellite images, and other inputs. The output became thousands of precise holdings, each with a small statistical edge. Some relations lasted briefly, and even the team could not explain why every relation worked.

The video calls Renaissance the ultimate computerized bee because it incorporates information from many parts of the world and distributes it through prices. The same information race can move at absurd speed. In 2013, a hacked Associated Press account published a false tweet about an explosion at the White House. The video says that algorithms moved the market by $136 billion within seconds, then reversed the move when they recognised the report as false. Hedge funds now buy satellite images to count cars in Walmart car parks before earnings, whilst Citadel pays Robinhood for order flow that reveals trades before they reach the market. The source uses these examples to show how far upstream traders now search for information.

The market portfolio and the cost of being clever

William Sharpe supplies the final step. Instead of judging traders one by one, he asks what they hold when their positions are added together. Every investor votes with money by choosing a combination of stocks. At any moment, all those holdings form one collective portfolio, which Sharpe calls the market portfolio.

The market portfolio earns the average return of all investors because it is the average of their holdings. It also avoids the fees and trading costs that active investors pay whenever they change position. The video therefore presents passive investing as a way to receive average performance before costs and better-than-average performance after costs. Active traders still perform the research that moves prices towards their collective value. The index investor receives the resulting portfolio without paying each researcher separately.

John Bogle turned Sharpe’s insight into an index fund in 1976. A fund can hold all 500 stocks in the S&P 500 in proportion to their market value, so changes in the collective judgement of investors update the holdings without a new decision from the fund owner. The video says that Bogle’s original fund beat roughly 90% of funds after its creation. It presents the result as a consequence of low fees rather than superior forecasting.

That is why Buffett could bet against the professional activity that made his own investing possible. Every trade gives an active investor another chance to be wrong, whilst the index fund keeps the collective market position and lets fees take their toll on the active side. The narrator also describes his own bankruptcy and three occasions when he had to borrow from friends during nine years in the business. The point belongs to the source’s personal account, not to a general measure of trading risk.

The closing example starts with 10,000 invested in the S&P 500 on 11 March 1942. A presentation in the video gives the later value as 51 million, with the investor having made one decision and left the portfolio alone. The video calls this the secret of winning at the world’s most sophisticated game: refusing to play. Its final joke is that a friendly stockbroker would have starved.

Limits

The video offers a clear explanation of why active trading can make passive investing work. It is also a narrated argument rather than an audit of the figures it quotes. The claims about the IBM block order, crypto trading volume, Coca-Cola’s returns, Bogle’s fund, the $136 billion reaction to the false AP tweet, and the$51 million S&P 500 result remain attached here to Art of the Problem. The captions contain several garbled names and numbers, so the note preserves the source’s figures whilst avoiding claims that would require a separate financial history.

The video describes index investing as a mathematical answer to the cost of active management. It does not discuss taxes, fund-tracking error, changes in market composition, inflation, a person’s time horizon, or the risk of holding an equity index. The source itself says that it is not investment advice. The passive-investing conclusion belongs to the argument about aggregate performance after fees. It is not a promise that an index fund wins over every period or suits every investor.

Further reading / references

• The Stock Market Always Wins, the companion article named in the video description.
• Benjamin Graham, The Intelligent Investor (1949).
• William F. Sharpe, “The Arithmetic of Active Management.”