Most people treat position size as a gauge of conviction — how much you believe in something. We treat it instead as a gauge of how much you could afford to lose if you're wrong.
There's a common intuition about position sizing that we think is quietly dangerous, and it's worth naming outright. The intuition holds that position size should track conviction: the more strongly you believe in something, the bigger the position you should take. It sounds reasonable, and it's how a great many investors actually behave. We think it gets the purpose of sizing backwards.
The trouble with sizing by conviction is that conviction isn't a reliable guide to being right. An investor's strongest beliefs aren't systematically their most accurate ones — if anything, the intensity of a belief can just as easily reflect how compelling a story sounds as how well-founded it is. Sizing by conviction ends up putting the most capital exactly where an error would be most costly, and it does so on the strength of a feeling with no guarantee attached to it.
We prefer to treat position sizing as a tool for controlling risk rather than expressing conviction. Under that framing, the relevant question isn't how strongly you believe in a position — it's how much you could stand to be wrong about it. A position should be sized so that if it fails, and any position can fail, the failure is survivable rather than ruinous. Size gets set by the tolerable cost of being wrong, not by how badly you're hoping to be right.
This has a consequence that a lot of investors find counterintuitive: it caps the size of even your best ideas. However persuasive a case looks, the chance that it's mistaken never disappears, and sizing by risk control means even the most attractive position is capped at a size whose failure you could absorb. That feels like leaving money on the table, and in the cases where the conviction was right, it is. It also keeps you standing in the cases where the conviction was wrong — and those cases aren't rare.
There's a subtler benefit here too, beyond risk control itself: this discipline tends to sharpen judgment. A position sized so its failure would be survivable can be assessed calmly, on the merits, because the stakes stay bounded. A position sized so large that failure would be catastrophic warps the thinking of whoever holds it, because admitting the mistake becomes too expensive to face. Bounding the size, oddly enough, often produces clearer thinking about the position itself.
This ties into a broader point about concentration and being wrong. Concentration expresses confidence in your own judgment, and confidence isn't the same thing as accuracy. An investor whose sizing tracks their confidence has arranged their capital so that the accuracy of their judgment determines whether they survive, and since judgment is fallible, that's a precarious setup. Sizing by risk control cuts that link — no single error, however confidently made, can wipe you out.
For an investor, the practical discipline is to ask, before putting on a position, not just how attractive it looks but how large a loss its failure would cause, and whether that's a loss you could genuinely absorb without damaging the broader plan. If the honest answer is that failure would be devastating, the position is too big, no matter how compelling the case looks. A stronger case doesn't change the arithmetic of what failure would cost.
We hold to this because the record of investors who sized by conviction and got it wrong is a long one, and because the whole logic of long-term investing depends on surviving your own errors long enough for your sound judgments to compound. Position sizing by risk control is, at bottom, about ensuring survival, and survival is the precondition for everything else. We size positions by what we can afford to lose, not by how much we're hoping to gain.