Conviction position sizing: how much to put in one stock

The last time you bought a big chunk of something, how did you decide on the size of the bet? Be honest. There is a very good chance the answer was some flavour of “I was really feeling it that day.” You read a cracking write-up, your heart rate went up a notch, and the number in the order box quietly crept higher than usual. That, my friend, is position sizing by dopamine. It is one of the most expensive habits in all of investing.

Position sizing is the boring name for the most important decision you make on every trade. How much to invest in one stock. Not whether the idea is good. How much. And almost nobody does it on purpose. I want to make the case that the size of a position should fall out of your conviction and your written plan, like the answer to a sum, rather than out of how excited you happen to feel at the moment your finger hovers over Buy.

Why we all size by feeling (and why it backfires)

The dopamine method has a cruel little twist built into it. Excitement is at its loudest exactly when it should be quietest. You feel most certain right after the price has run and the bullish thread has done its work, which is to say right at the top of your own emotional cycle. So sizing by feeling loads you up heaviest at the worst moments. It also leaves you timid precisely when something cheap and unloved is the actual opportunity. Your enthusiasm is a contrarian indicator wearing a very convincing disguise.

There is a second problem, and it is even sneakier. Feelings do not leave a record. When a feeling-sized position goes wrong, you cannot examine the decision, because the decision was never written down. A book full of bets you cannot reconstruct is a book you cannot learn from, and learning is the entire point of doing this for years rather than once.

Conviction is a thing you can write down

So reframe it. Conviction is a claim about the world, and a claim can be written down and checked later. That is what a thesis is for. In Strata a thesis is the home for the actual reasons you own something. You write down the drivers you think will play out and the assumptions underneath them, the invalidation signals that would prove you wrong, and a conviction score that says in plain numbers how sure you are.

Once conviction is an explicit score rather than a vibe, sizing has something real to attach to. A position you would defend in a courtroom, with drivers you can name and a clear sense of what would change your mind, earns more of your book than a “might run” punt. Not because a formula told you so, but because you decided that on purpose, in advance, when your head was cool. The high-conviction idea and the speculative flier should not be the same size, and yet sized by feeling they very often are, because on the day you bought each one you felt roughly equally great.

This is the same logic behind tiering rather than treating every name as equal. A small set of ideas you understand deeply should do the heavy lifting, and the long tail of maybes should stay small enough that being wrong about any one of them is a shrug, not a wound. The aim is not to be right about everything. The aim is to be big where you are right and small where you are guessing.

Decide the size before you buy, not after

The trick that makes all of this work is embarrassingly simple: write the size down before you place the trade. Not the share count. The intent. Roughly what slice of the book this idea deserves, given how it scores against everything else you already own, and why.

Doing it in advance defuses the dopamine. The number you write when you are calmly comparing one idea against your other holdings is a far better number than the one you would type in the heat of a breakout. It also forces an honest conversation with yourself. If you cannot justify a large size in writing, that is useful information. The plan was never as strong as the feeling.

A written sizing plan also turns into something you can grade later, which is where the real compounding happens. Your journal records the plan against what actually occurred, including the realised R on the trade, so you find out whether your big bets were the ones that paid and your small ones were the ones that flopped, or whether, awkwardly, it was the other way round. If you have been reading along, you will know this is the same discipline loop we keep banging on about: write the reasoning down, then confront it with what happened. Sizing is just the part of the loop most people skip.

None of this is a recommendation about what your sizes should be. There is no magic per cent, and anyone selling you one is selling you a formula to follow blindly. What matters is the framework. Link the size to a written, scored thesis, decide it before you act, then review it after. What the actual numbers are is yours, and depends on things only you know.

Was your conviction actually any good?

Almost nobody asks this one. When you felt sure, were you right to? Conviction is only worth sizing on if it is calibrated, and the only way to know is to check your own track record. This is the part that separates investing from gambling with extra steps.

Say, purely to illustrate with round numbers, that you sort a year of trades by the conviction score you gave each one going in. If your high-conviction names genuinely outperformed your low-conviction ones, your conviction is calibrated, and sizing by it is rational. If they did not, that is gold, because it means your sense of certainty is noise, and you have just learned to stop trusting it and to size flatter until your judgement improves. Either answer makes you better. The only losing move is never looking.

A thesis with invalidation signals helps here too. When a signal breaches and the thesis goes to AT-RISK, you get a sell prompt, but you also get a data point about your original conviction. An idea you called high-conviction that keeps tripping its own invalidation conditions was never high-conviction. It was high-confidence, which is a completely different and far more dangerous thing. This is the same machinery that decides when to sell on a broken thesis, pointed back at the entry.

The boring discipline that beats the dopamine

The whole thing fits on a postcard. Size by what you can defend, not by how you feel. Write the conviction down as a thesis with real drivers and real invalidation signals, and let the size follow from that score and from how the idea stacks up against the rest of the book. Commit to the size before you trade, journal it, then check afterwards whether your conviction was calibrated and adjust how much you trust your own certainty next time.

It is slower than feeling it. It is also the difference between a book you can reason about and a pile of bets you happened to make on good days. Strata exists to make conviction and sizing explicit and reviewable instead of a private feeling you cannot inspect. If a process you can actually grade sounds better than a gut you cannot, you can build your portfolio and start sizing on purpose. Bring your own conviction. We will help you keep it honest.