Writing Down The Thesis And What Would Kill It

Before entering a trade, write two things down: why you think it will work, and what specific, checkable fact — not a price level alone, but a piece of information — would tell you that reasoning was wrong. The first part is the thesis. The second is the invalidation condition, and it's the part traders skip most often, because writing "I could be wrong about this" feels uncomfortable at the exact moment you're most confident.

The reason this habit matters is that markets are very good at producing evidence that feels meaningful in the moment but is actually noise, and a trader without a pre-written invalidation condition ends up deciding, in real time and often under stress, whether each new piece of price action means "hold" or "exit." That decision, made after the fact and often after a loss has already started, is exactly the situation where cognitive biases are strongest — the desire to be right, the sunk cost of time already spent on the idea, the discomfort of admitting a mistake. A thesis and invalidation condition written down beforehand, when the trader had no emotional stake in the outcome yet, is a much more reliable guide.

Worked example

A trader buys a stock on the thesis that a new product launch will drive a meaningful revenue acceleration starting next quarter, expected to become visible in the following earnings report. The invalidation condition, written at entry: "if the company's own guidance on the next earnings call does not mention accelerating demand for the new product, the thesis is wrong regardless of what the stock price has done in the meantime." Two months later the stock is down 8% on no company-specific news — a level move that, without the written thesis, might trigger a panicked exit. But nothing in the invalidation condition has happened yet; the earnings call hasn't occurred. The trader holds, per the plan, distinguishing a price move that doesn't touch the thesis from one that would. When the earnings call arrives and guidance does confirm accelerating demand, the position is held through the noise it would otherwise have been shaken out of.

The invalidation condition has to be a fact, not a price. "Exit if it drops 10%" is a stop-loss, which is a separate and useful risk control, but it isn't the same as asking whether the reason for the trade is still true — a stock can drop 10% on broad market weakness with the original thesis completely intact, or sit flat while the thesis quietly breaks.