Position Sizing: Why We Pour Quarter-Kelly

"How much should I put on this one?" is the question that gets people in more trouble than almost anything else at the bar or in the market. Too little and a good idea barely matters. Too much and one bad night wipes out a month of good ones. There's actually a well-known answer to this problem, and then there's the practical version of that answer that real humans can actually live with. We pour the second one.

The idea behind Kelly sizing

Decades ago, a mathematician figured out a formula for how big a bet should be, given how good your edge is and how confident you actually are in it. The intuition is simple even if the math isn't: a stronger, more confident edge earns a bigger pour, a weaker or shakier one earns a smaller pour, and a real bartender would never size two very different situations the same way. That's "full Kelly" — the mathematically optimal bet size for maximizing long-run growth.

Why "mathematically optimal" isn't the same as "livable"

Here's the catch nobody tells you about full Kelly: it's optimal for a calculator, not for a person. It produces genuinely brutal swings — the kind of drawdowns where your account can lose half its value even when the underlying edge was perfectly real and the strategy was working exactly as designed. Almost nobody has the stomach to sit through that without panicking and abandoning the plan at the worst possible moment. And an edge you abandon halfway through isn't an edge anymore, it's just a loss you gave up on too early.

Why a quarter pour instead of a full one

So we scale it down to a quarter of that size. You give up some of the theoretical long-run growth rate — that's the honest trade-off, we're not going to pretend it's free — but you get a dramatically smoother ride in exchange. The swings shrink a lot faster than the growth rate does, which is a genuinely good trade for a real person trying to stay in the game for years, not just survive one lucky quarter. It's the difference between a pour that gets you through the whole night in good shape and one that gets everybody cut off by ten.

This is also why sizing isn't something we hand you as a suggestion and hope you do the math right under pressure. It's built into how every recommendation gets shown to you — scaled to the edge and the confidence of that specific call, the quarter-Kelly way, every time, automatically.

The bottom line

Bigger isn't always better, and "optimal on paper" isn't the same as "survivable in practice." We'd rather hand you a sizing approach you can actually stick with through a bad week than one that's theoretically perfect and practically unbearable.

See how sizing shows up across our different strategies on the menu, or head back to the blog.