Why Linear Thinking Fails in Nonlinear Markets: An Investment Review

The linear-thinking trap in investing starts when a market is treated like software with a stable “input A produces output B” relationship. Single-cause stories, historical fitting, and confidence in a clean model can make a thesis look rigorous while hiding the variables it cannot explain.

This review was written on January 20, 2026, the same day as the Kunlun Tech investment review. That concrete case makes the problem easier to see: even when an AI-business thesis has not been falsified, the price path is still shaped by execution timing, valuation, risk appetite and position pressure at the same time. Treating any one of those variables as a stable input makes the model look more deterministic than the market really is.

Why engineering habits can over-linearize markets

In software, inputs, dependencies, and runtime conditions can often be controlled. Markets combine earnings, valuation, liquidity, policy, competition, expectations, and positioning. The same positive event can produce different price outcomes under different starting conditions. Historical repetition can suggest a hypothesis; it cannot create a deterministic law.

The more dangerous form of overfitting appears when every counterexample is absorbed by adding another explanation. A model that can explain every outcome after the fact can no longer be falsified.

Falsification conditions matter more than another explanation

I now want each thesis to state what evidence would prove it wrong. If the case depends on order growth, define which order, revenue, or cash-flow change breaks the assumption. If it depends on an industry cycle, track the opposite signals in pricing, inventory, and competition. When a falsification condition is triggered, reduce risk first instead of immediately inventing another story.

Position size is also an expression of uncertainty. A thesis with many unresolved variables should not be converted into unlimited exposure merely because the long-term narrative is attractive. Related reviews are collected in Investing. For execution under stress, see the Trading Discipline Review. For outcome bias after selling, read the Sold Too Early Review.

A rule set that makes the model testable

Before entry, keep separate lists of supporting evidence and falsifying evidence. Set a position limit instead of turning the thesis into a belief. During review, classify the failure as an information error, model error, or execution error. None of this makes the market predictable, but it makes mistakes easier to detect and limits the damage from one wrong model.

This article discusses an investment decision framework, not a stock recommendation.