On September 25, the S&P 500 Index closed just 0.7% below its August 13 record high. Yet only 24.5% of the stocks in the Multi-Cap universe we follow—U.S. common stocks—were above their 50-day moving averages, and only 43.1% were above their 200-day moving averages. That marked the first time since the Multi-Cap dataset began in 1981 that the S&P 500 was within 1% of an all-time high while less than 25% of stocks were above their 50-day and less than 45% were above their 200-day. The question is: Does this mark the end of a stealth correction, or is it an erosion that is more serious longer term? It could be both. If we get more of the same in the form of a year-end rally before a bear market ultimately unfolds, we could look back at this as the writing that was on the wall. Either way, the fact remains: breadth is very poor. That is why we have such a selective market, with the major indexes being driven by a relatively small list of capitalization-weighted names. minerviniprivateaccess.com

Ahem, it’s “Superior Intelligence” now

Even though I know this won't satisfy everyone. I'm hoping this will at least make enough sense that those of who are intelligent and truly objective will have a chance at seeing the reality of markets and pricing efficiency. Let’s take the AI argument all the way to a logical conclusion. The claim is that AI can process more information than any human, analyze every relevant variable, eliminate emotion, learn from enormous amounts of historical data, test millions of possibilities, and ultimately make better trading decisions than a human ever could. Fine. For the sake of argument, let’s assume all of that is true. It's perfect, or at the very least, much better than us as humans trading. Let’s assume you can give AI the proper information, the proper objectives, the proper constraints, and enough data, and it can optimize a trading strategy to near perfection. No fear. No greed. No hesitation. No fatigue. No ego. No emotional mistakes. Just mathematically optimal decisions. Now ask yourself the obvious question: Who wouldn’t use it? If AI really provides that kind of advantage, every serious professional trading operation is going to use it. Hedge funds, proprietary trading firms, market makers, institutions—anyone responsible for a meaningful amount of trading volume—is going to employ increasingly sophisticated algorithms and AI models. And that creates a paradox many refuse to acknowledge. If everybody has access to increasingly intelligent systems, and those systems are all analyzing essentially the same market, the same prices, the same volume, the same economic data, the same earnings reports, the same news, and increasingly similar alternative datasets—what happens when they all become “perfect”? Where does the advantage come from? If two tennis players were literally identical in skill, speed, strength, strategy, stamina, and execution—and neither one ever made an error—how does one consistently beat the other? He doesn’t. It would be a tie. Trading is no different. For someone to generate alpha, there must be some form of inefficiency, informational advantage, analytical advantage, tactical superiority, behavioral advantage, execution advantage, or simply a difference in how information is interpreted. I've heard comparisons made to chess. Markets are not chess. Not even close. Chess is a closed system. Which makes it a terrible analogy for comparison. The rules are fixed. The board is visible. The pieces are known. The possible moves are defined. There is no surprise inflation report halfway through the game. The queen doesn’t suddenly announce an earnings miss. Interest rates don’t change while you’re deciding whether to move your bishop. There is always a right and wrong move. The stock market is an open, adaptive, probabilistic system operating on incomplete information. There often is no single mathematically “correct” move because the answer depends on an outcome that has not happened yet. AI can calculate probabilities faster than I can. It can process more information than I can. It can recognize relationships I might never see. But it cannot turn uncertainty into certainty. And here is the ultimate contradiction in the “AI will perfect trading” argument: The more universally effective a trading edge becomes, the more capital attacks it—and the more that edge gets arbitraged away. Markets adapt. The whole argument against breakout trading is it doesn't work as well as it did because too many people are using it. Well, why wouldn't that same argument be made for AI, which ultimately would be applied in a much more universal, consistently predictable manner than humans applying breakout trading. Even Jim Simmons understands this, and that understanding is why he is successful. That's why he as admitted to manually monitoring and changing his models routinely. And that humans can override his models at the risk level. If humans are constantly testing, changing and overriding the AI models - based on their findings - then AI is a tool, not a solution, and certainly not autonomous. In his TED interview, Jim Simons said: “These things fade after a while; anomalies can get washed out. In another interview, Simmons confessed: “The computer is just a tool that we use.” I rest my case. minervini.com minerviniprivateaccess.com

If AI could truly figure out how to trade, how would one AI trader beat another? IT COULDN’T. You need inefficiencies to generate alpha. If every AI sees the same data, finds the same patterns, and makes the same “optimal” trades, the edge gets arbitraged away. Even Jim Simmons understands this. That's why he as admitted to manually monitoring and changing his models routinely. And that humans can override his models at the risk level. If humans are constantly testing and changing the AI models - based on their findings - then that AI is a tool, not a solution. In his TED interview, Simons said: “These things fade after a while; anomalies can get washed out. In another interview, Simmons confessed: “The computer is just a tool that we use.” I rest my case.