Complex Adaptive Systems (Stonk Market) and How to Beat Them

Source: Complex Adaptive Systems (Stonk Market) and How to Beat Them, Benjamin, 18:06, uploaded 2021-06-11, Watch Later position 393.

An ant colony contains simple creatures that follow local rules, yet the colony builds paths, fungus gardens, and rubbish pits without a government or central plan. Benjamin uses that puzzle to approach the stock market. The individual traders may make understandable decisions whilst their interaction produces a system that changes its own conditions, absorbs information, and defeats the models that try to predict it.

From ants to Yellowstone

The colony is the opening example of aggregation. One ant forages, another stacks sand, and a third appears to avoid pulling its weight. Each ant has a narrow range of possible actions. The colony acquires properties that no single ant possesses: it moves as a whole, responds to its surroundings, and follows a life cycle that resembles an organism. The question is how independent decisions organise themselves into a coordinated system.

The Yellowstone wolf story gives the same problem a larger scale. From 1870 onwards, park employees hunted wolves in an attempt to remove predators. By 1926 the last wolf pack had gone, and by 1995 deer had spread across the park and eaten down its vegetation. The US government released fourteen wolves. As the wolf population grew, deer avoided riverbanks and valleys that had suffered from overgrazing. Vegetation recovered, beavers returned to the rivers, and populations of mice, hawks, foxes, and bears expanded.

The trees changed the system again. Their roots held the soil, reduced erosion, narrowed river channels, and in some places changed the path of the river itself. Benjamin’s point is the scale of the result. A small change in the conditions of a system can produce a large change in its behaviour, even when the individual animals remain fairly easy to understand.

Why the parts fail to explain the whole

Suppose we can predict a deer’s actions with roughly 70 percent certainty. We then use that prediction inside a larger model that also includes wolves, vegetation, and the other animals. The next prediction rests on the first one, which rests on an assumption about the starting conditions. Each layer adds uncertainty. A detailed account of every part still leaves the system’s combined behaviour out of reach.

The stock market supplies the financial version of the same problem. A basic model gives each independent trader one aim: make as much money as possible. Every purchase needs a seller, though, and the counterparty may be an institution that has better information, a different risk tolerance, or a model that has already priced in the event that the individual trader just discovered.

Benjamin invents a Microsoft example around Bill Gates’s divorce. An institution might use satellite images to track changes in Melinda Gates’s behaviour, fit those changes to a linear regression model, consult behavioural psychologists, and speak to Microsoft insiders. The institution could decide that a divorce is likely months before public news appears. A retail trader sees the news after it breaks and assumes the market has yet to respond. Both trade Microsoft, while their information and assumptions occupy very different positions.

That difference makes the price move. People bring different information, bias, and tolerance for risk into the same exchange. A clear understanding of each participant leaves the collective system unresolved because the participants respond to one another.

Aggregation, emergence, adaptation, and nonlinearity

Benjamin jokingly calls the four ideas that follow the “you won’t principles”. Aggregation describes the ant carrying one piece of sand. Emergence appears when thousands of such tasks produce an organised colony with properties that no task contains by itself.

Adaptation makes the market difficult to beat. Every participant learns from the environment and changes their position. Benjamin imagines a system that predicts market corrections with consistent accuracy. Its owner could borrow heavily and buy deep out-of-the-money puts, following the example of Bill Hwang. The strategy might work for a while. As other traders watch the option volume, demand changes implied volatility and the market incorporates the apparent edge. A successful prediction changes the market that made the prediction possible.

Nonlinearity explains why the change can become hard to trace. A car follows equations for acceleration, friction, and maintenance, so its movement remains predictable within the model. Benjamin places the market in a different class. A small input can produce a large-scale change whose path back to its initial cause disappears.

Momentum trading shows the mechanism. Tim Cook unveils an Apple car that looks better than expected. Institutions buy Apple shares, algorithms detect the price movement, and retail traders see the story on WallStreetBets. They buy out-of-the-money options, which raises gamma across the option chain. Market makers then buy more shares to hedge their exposure, and those purchases continue the rise. The market’s response becomes a feedback loop whose later stages strengthen the first movement.

The efficient market hypothesis offers a cleaner answer. If risk and reward are symmetrical, investments have an expected value of zero and nobody can achieve superior trading results with consistency. Benjamin explains the idea through a loaded-sounding dice game. A roll from one to five earns ten dollars, whilst a six costs fifty. The five winning outcomes make the game look attractive to someone who knows basic statistics. Repeated rolls eventually move the result between small gains and losses, bringing the value towards zero.

Efficiency has limits

The video pushes the efficient-market idea towards absurdity. If the market prices every photon hitting the Earth, the local weather that follows, and the effect on wheat futures, Goldman Sachs would need analysts on the Moon to price asteroid impacts, tides, salmon migration, and Atlantic salmon futures. The joke makes room for a qualification: real institutions face portfolio mandates, liquidity needs, compliance, and Federal Reserve rules. Their trading cannot respond to every possible fact.

The Zoom episode shows how a public market can misallocate capital. During the pandemic, shares in a company called Zoom rose by thousands of percent. The company was Zoom Technologies, a Chinese business that had been out of operation for years, rather than the video-conferencing company Zoom Video Communications. The SEC stopped trading because investors had confused the two companies. The incident sits badly with a picture of rational investors pricing every asset correctly.

Benjamin’s solution is diversity of opinion. People who challenge one another’s ideas reduce the space in which a shared mistake can grow. Each participant may hold incomplete knowledge, yet the exchange forces people to test their positions against other arguments and update them when new information matters. That process can make a public exchange efficient enough to leave little opportunity for an ordinary trader.

The same mechanism shows where inefficiency can appear. “The housing market will never go down” is one example. “GameStop is a dying company” is another. The video ties the GameStop case to its extreme short interest and to a market in which almost everyone appeared to share the same bearish view. A consensus can carry bias into prices, which creates the possibility of a large move when a small change disturbs the common story.

GameStop and the limits of hindsight

Nobody could predict the full GameStop episode. Its emergent properties included people putting GameStop on billboards around the country. One can trace a chain through the initial due diligence, posts on WallStreetBets, new members joining after mainstream attention, and the later public spectacle. The chain still fails to identify the cause of the billboard. Any one event, a combination of events, or an unrecognised event could have supplied the change. If the initial due diligence had never reached WallStreetBets, the billboard might still have appeared through another route, or it might never have existed.

The episode looks like a great opportunity in retrospect. The bearish case may have deserved its pessimism, and most similar stories will move the other way. The asymmetry lies in the distance between the possible loss and the possible gain. Benjamin places that asymmetry where a large group cannot imagine how it could lose. The advice remains a way to inspect consensus rather than a method for forecasting the next GameStop.

That leaves the opening question partly unresolved. Complex adaptive systems offer useful language for aggregation, emergence, adaptation, feedback, and nonlinearity. They explain why a market can contain many intelligent participants and still generate outcomes that nobody can reconstruct. They do not give Benjamin confidence that the theory will predict real events, and he ends by admitting that he would not rely on his own advice for managing risk or earning outsized returns.

Limits

The source is a comic explanation rather than a documented financial study. Its claims about Yellowstone, the Microsoft scenario, Bill Hwang, the Zoom Technologies halt, option gamma, and GameStop remain attached to Benjamin. The video gives no citations or references for its figures and compresses some market mechanics for humour. The captions also contain a few garbled words and censor placeholders.

The argument identifies conditions under which consensus may create an opportunity. It does not provide a tested trading strategy, a probability for finding such an opportunity, or investment advice. The final joke about playing Ellie Goulding in the car makes the same point in miniature: theory can describe a system whilst leaving the person who describes it unable to predict his next decision.