🧐 Most of our theses aren't any good

My glorious return to a second year of active trading has taught me a valuable lesson in only twenty-seven days. I came back from 2025 wholly unprepared for the emotional temperament and intellectual rigor necessary for trading. A nine-month absence atrophied my critical thinking, unbeknownst to me, such that I re-entered with the naïveté shared by most everybody else. The quintessential quality of this herd naïveté is self-deception and cognitive bias errors, believing in one's flawed process and reasoning. While I end three weeks with a +20% portfolio return, I scrutinize my own failures and recognize that I simply got lucky again. The overwhelming majority of my theses were not any good; they were dangerous.

Let's start with a range of theses. First, you can have none. This is gambling. My opinion on gambling, it is generally immoral, spiritually bankrupt and ruinous, and provides certain financial destruction. On where to gamble, I used to believe the markets are probably the best venue. I thought expected value of gambling on 0 DTE exceeds the best house casino games such as blackjack, that 0 DTE is essentially a 50% probability coin-flip with less vig. If a .45 caliber were pointed to your head, and you had to double your money immediately or lose your life, I believed the best place to go is the tape. I have since changed this position, arguing that 0 DTE gambling is as terrible as a house casino, due to theta decay, implied volatility crush, and spread. It turns out there is simply no safe or fair place to double your money instantly. I cannot say much more about a thesis that does not exist.

First, comparatively yet perhaps more dangerous than gambling, and most common amongst the public, is the thesis that what happened yesterday predicts what happens today. My opinion is this cognitive error is the foundation of technical analysis. The public thinks of a self-professed "daytrader" as being a technical analyst, right in that this analyst is statistically the largest majority of known investors. A technical analyst loves ticks, patterns, and shapes. They arm themselves with a dramatic lexicon: "coil up", "break-out", "support". They dream of mean reversion. These folks dominate social media and live streaming, sell books and courses veneered in complex numbers and formulas, disarming the audience with presentations of pretty shapes.

What is most dangerous about technical analysis is that the human brain is wired to perform such pattern recognition. It is an evolutionary process necessary for survival. Applied to investing, it is fatal. An investor must not be engaging in stylistic technical analysis to fall prey to this bias. After scrutinizing my own recent theses, I see this cognitive error in virtually all of my reasoning. My best hope to overcome my limitation is to repeatedly observe, over and over, that I am performing it. With constant self-discovery, I can eliminate my likelihood to fall prey to the bias that the past predicts the future.

The second tier of useless theses is making inferential connections that could possibly make sense, without a real signal. For example, one can reason that something over there has a connection to something over here. It could be direct, or it could be degrees-of-separation. Others may establish these connections in the public narrative, or you may invent it yourself. The difficulty of overcoming this useless thesis arises from its seeming logical consistency. This error is a form of logical bias where the mind has produced a cause-and-effect, that while idealistically true in some Platonic form, nonetheless fails when applied to the conflicts and dynamics of the real world. This failure can be catastrophic, or it may produce winnings, and be incorrectly attributed as evidentially correct. The way to escape this type of thesis is to consider all the other factors the logical chain has excluded. Then, one can place a real possible probability that each link in the chain is correct and complete. After discovering what-is-left-out, and ripping apart each link in the chain, the probability of success of seemingly logically consistent conclusions can be reduced to far less than 1%, trending to zero, or without any practical correlation.

The third tier of useless theses is using facts that you know of a certain domain to falsely establish your understanding of that domain. For example, you may have learned ten, a hundred, or even a thousand facts about a domain, but you may grossly misunderstand that domain. I believe the more you learn about a domain, the more likely you are to use those facts towards incorrect understanding. The question is then proposed: "so how do you know that you understand a domain?" I will have to think about this more, but at the time of this writing, I believe this may be an impossible question to imperatively answer. One can only be extremely cautious and self-scrutinizing, recognizing one's intellectual humility, and discovering one's small place in the world. If we raise our own bar for what constitutes a subject-matter expert, we can eliminate the increasing probability that we produce this form of useless thesis.

Threads holding these useless theses together:

Desire or pressure to win

A cryptographically signed ticker that produces mostly interpretive noise

Apophenia (the human tendency to perceive meaningful patterns or connections in random, unrelated data or events)

Narrative fallacies

Distinguishing accumulated facts from structural understanding

What does make a good thesis? We imagine Munger might say:

"I think active trading is an idiot's game. Why would anyone with half a brain spend their finite life trying to out-guess random tape wiggles when you could compound capital in wonderful businesses run by honest people? Your mind is far too sharp and clear-eyed to waste on short-term market noise. Take this exact epistemic rigor, throw away the minute charts entirely, and apply it to multi-year business fundamentals within a defined circle of competence."