Every few years the financial industry rediscovers artificial intelligence as the solution to the problem of beating markets. The tools change. The promise stays the same. And the fundamental error underneath it all remains unchanged: the assumption that more data, more variables, and more computational power will eventually crack the market’s code. Trend following has a different answer, and it has been producing results for decades.

Business 2.0 featured an article by Carla Fried that raised the argument of artificial intelligence as a potential investment tool. Many of the points and analogies presented in the article actually assist in explaining trend following’s success. Here are excerpts from the article with our commentary.

Silicon vs. Gray Matter: The Wrong Question

If ever there were a field in which machine intelligence seemed destined to replace human brainpower, the stock market would have to be it. Investing is the ultimate numbers game, after all, and when it comes to crunching numbers, silicon beats gray matter every time. Nevertheless, the world has yet to see anything like a Wall Street version of Deep Blue, the artificially intelligent machine that defeated chess grand master Gary Kasparov in 1997. Far from it, in fact: When artificial-intelligence-enhanced investment funds made their debut a decade or so ago, they generated plenty of media fanfare but only uneven results. Today those early adopters of AI, like Fidelity Investments and Batterymarch Financial, refuse to even talk about the technology.

Data flows in not just from standard databases but from everywhere: CNN, hallway conversations, trips to the drugstore. “Unless you can put an emotional value on certain events and actions, you can’t get the job done.” Naturally, investors don’t process this hodgepodge of inputs according to some set of explicit, easily transcribed rules. Instead, the mind matches the jumble against other jumbles stored in memory and looks for patterns, usually quite unconsciously. “Often, great investors can’t articulate the nature of their talent. They’re like pool players who make incredible trick shots on intuition.” Fine for them, but how do you code that?

TurtleTrader® comment: Coding emotions? Measuring value? Why not go right to the market price to measure value like trend followers do every day? Why make the problem more complicated than need be?

The article identifies the real problem without realizing it. The effort to code intuition, to quantify hallway conversations and trips to the drugstore, assumes that the information advantage great investors possess is something other than price. But price is the final output of every piece of information every market participant holds. It is the aggregate of all of it: the CNN headlines, the institutional research, the insider sentiment, and the crowd psychology. Trend followers skip the inputs entirely and read the output directly. The price already knows everything the AI is trying to discover. Going to the source is not a simplification. It is a more precise form of analysis than any multi-variable model can achieve.

Fair Value and the 18-Variable Problem

Andre Archambault, for example, manages Standard & Poor’s Neural Fair Value 20, an AI-enhanced model portfolio open to subscribers of S&P’s Outlook newsletter. His AI software analyzes the 18 financial variables that he uses to calculate fair values for his universe of 3,000 stocks.

TurtleTrader® comment: Instead of using 18 variables to arrive at fair value, why not accept the one variable, the market price, that actually is the fair value? No calculation needed!

This is one of the clearest expressions of the trend following philosophy in contrast to fundamental and AI-driven approaches. Fair value is not something you calculate from 18 inputs and then compare to the current price. The current price is the fair value. It is the only number that every buyer and seller in the market has agreed upon at this moment. It reflects their collective assessment of all 18 variables the model is trying to measure, plus every other variable those 18 leave out. A model that takes inputs and arrives at a number different from the current price is not revealing hidden value. It is betting that its model is smarter than the market’s aggregate judgment. That bet rarely pays over time. For the full framework of how trend following uses price as its only input, see the TurtleTrader rules.

The Problem With Every Quantitative Approach Built Around Prediction

History shows that [problems] eventually strike every quantitative investment approach, from blunt rules of thumb like the “dogs of the Dow” (buy the 10 highest-yielding Dow stocks each year) to the arcane strategies cooked up by Ph.D.s at hedge fund D.E. Shaw Group. The reasons are endless: Financial conditions can change; other investors can catch on, eliminating a winner’s edge; or tastes can shift, and what excited the market.

TurtleTrader® comment: If you are trying to foresee “anything,” no strategy will work. If you are trying to predict tomorrow, that will be a bust too. Arcane trading strategies? Stop. Will not work. Using a computer to tell you about tomorrow? A sure fire plan to lose money. From this moment forward when reading any financial press keep track of the prediction mentality that seems so pervasive. Does anyone in the mainstream ever acknowledge that the very notion of prediction is a false idol?

The article frames this as a technical problem: strategies decay because conditions change, or because other traders arbitrage away the edge. But the deeper issue is that every approach built on prediction carries a fatal flaw from the start. Prediction assumes the future is knowable from current data. Markets exist precisely because participants disagree about the future. The moment a predictive model’s assumptions become widely known, the market adjusts to price them in, eliminating the edge. There is no stable predictive model because the market is not a static phenomenon waiting to be decoded. It is a continuously adaptive system made up of human beings changing their behavior in response to what they observe.

Trend following does not predict. It reacts. It has no view on where a market will be tomorrow, next week, or next year. It has a view on what the market is doing right now, and a set of rules for responding to that. Because it makes no prediction, it cannot be arbitraged away by other participants figuring out its forecast. The rules can be published in a newspaper, as was famously noted about the original TurtleTrader system, and they would still work, because the edge is not in the secret of the entry signal. It is in the discipline to follow the system through drawdowns, to size positions correctly based on risk management principles, and to let winners run without interference. Those behaviors are psychologically difficult. They cannot be arbitraged away because most people will not do them consistently regardless of how well they understand the logic.

The Prediction Mentality in Financial Media

The TurtleTrader commentary on the Business 2.0 article ends with a challenge worth taking seriously: from this moment forward, when reading any financial press, keep track of the prediction mentality that seems so pervasive. Does anyone in the mainstream ever acknowledge that the very notion of prediction is a false idol?

The answer is almost never. Financial media is built around prediction. Forecasts, targets, outlooks, and estimates are the product the industry sells. Admitting that prediction does not work as a basis for trading would undermine the entire apparatus of analyst recommendations, quarterly guidance, and market commentary that generates fees, subscriptions, and airtime. Trend following’s success is, in part, a direct consequence of operating outside that apparatus entirely. It trades price. It acknowledges uncertainty. It builds rules that work across conditions precisely because they do not depend on knowing what those conditions will be. For the full story of how this approach was built, tested, and proven, see the TurtleTrader story and the broader trend following framework.

Frequently Asked Questions

Why have AI-enhanced investment funds produced uneven results?

Because they are still built around prediction. Whether the inputs are 18 financial variables, neural networks trained on historical data, or sentiment analysis from news sources, the underlying assumption is that the future can be forecast from current information. That assumption is what fails. Markets are adaptive systems made of human participants who change their behavior in response to what they observe, making any predictive model’s edge temporary at best.

Why is market price the best measure of fair value?

Because it is the only number every buyer and seller has agreed upon at this moment. It reflects the collective judgment of all market participants, incorporating every variable a model might try to measure plus all the ones it leaves out. A model that calculates a different number from the current price is betting that its formula is smarter than the market’s aggregate assessment. That bet has a poor long-term record.

Why doesn’t trend following suffer the same decay as other quantitative approaches?

Because it does not predict. Every approach built on prediction eventually fails when conditions change or other traders arbitrage away the edge. Trend following reacts to current price action with no forecast of future prices. Its edge comes from the discipline to follow the system through drawdowns, size positions correctly, and let winners run, behaviors that are psychologically difficult and cannot be arbitraged away.

What is the prediction mentality in financial media?

The pervasive assumption, built into analyst recommendations, quarterly guidance, and market commentary, that the future can be forecast and that having the right forecast is the key to investment success. Trend following rejects this assumption entirely, operating instead on the principle that the market’s direction is unknowable in advance and that the edge comes from systematic response to what the market is actually doing.

How does trend following use price differently from AI investment models?

AI models use price as one input among many, comparing it to a calculated fair value derived from other variables. Trend following treats price as the only input needed, on the basis that price already reflects all available information. The question is not whether the price is right relative to some model. The question is which direction the price is moving and whether that movement is strong enough to hold a position.

Trend Following Systems
Want to learn more and start trading trend following systems? Start here.