How much to risk

The Math of an Edge

The size that grows fastest, then half of it

There's a mathematically optimal fraction to bet given a known edge — the Kelly criterion — and it turns out to be far more aggressive than anyone can psychologically tolerate. Two things go wrong in practice: you never truly know your edge, and overestimating it makes full Kelly catastrophic rather than merely bumpy. So practitioners use a fraction of it, and most retail traders land somewhere near 1–2% of the account per trade. Not because that's optimal, but because it survives being wrong about how good you are.