Sizing anchored to your worst trade
Most position-sizing rules pick a flat percentage — risk 1% or 2% of the account per trade — and never ask whether that number is actually the one that grows equity fastest. Optimal f does ask, and it answers with maths rather than a round number. Given the full shape of your results — how often you win, how big the wins are, how big the losses are, and crucially how big your single worst loss was — there is exactly one fraction of equity that maximises the geometric growth of your account over many trades. That fraction is optimal f, and the tool above finds it by testing every level and picking the one with the highest expected log-return.
The reason optimal f keys off the largest loss is subtle but important. Geometric growth is unforgiving of big drawdowns: a single trade that loses a large fraction of the account permanently lowers the base that everything after it compounds from. So the worst loss you have actually taken sets a hard ceiling on how large you can size before that one kind of trade starts eating the compounding. Feed in a nastier worst loss and optimal f drops immediately — the formula is telling you the tail risk in your record does not support bigger bets. That is a more honest signal than a flat percent that ignores your history entirely.
Why nobody sane trades full optimal f
Optimal f maximises growth, but growth and comfort are not the same thing. At full optimal f the drawdowns are savage — peak-to-trough falls of 60% to 95% are routine, because the fraction that grows fastest is also the fraction that swings hardest. Worse, the curve is asymmetric: sizing a little below optimal f costs you only a sliver of growth, but sizing a little above it costs growth and piles on risk, and sizing well above it turns a winning system into a losing one. That is why the practical use of optimal f is as a ceiling to stay under, not a target to hit. Half optimal f typically keeps most of the growth while roughly halving the drawdown; quarter optimal f is calmer still and is where a lot of professional money actually sits. The table above shows how flat the growth is on the left of the peak and how fast it falls on the right.
How to use it
1. Enter your win rate and your average winning and average losing trade in dollars — pull these from your trade log, not from memory.
2. Enter your largest single losing trade. This is the input that makes optimal f different from a flat percent; be honest, and if you expect a worse loss than you have seen, use that.
3. Read optimal f and the f$ figure — the equity that should back each one contract or position unit. Trade one unit per f$ for full optimal f, one unit per 2×f$ for half f, and so on.
4. Almost always, size at half or quarter optimal f. The verdict and table show what you give up (very little growth) and what you gain (far smaller drawdowns).
Common mistakes
Trading full optimal f. It is a mathematical ceiling with brutal drawdowns, not a comfortable setting — dilute it. Understating the worst loss. Optimal f is only as safe as the largest loss you feed it; slippage, gaps and liquidations produce losses bigger than your stop, so use a realistic worst case. Using too few trades. Optimal f computed from a handful of trades overfits to luck; recompute as your sample grows. Forgetting it assumes independence. If your losses cluster — correlated positions, the same setup failing in a regime — real drawdowns run deeper than the fraction implies, so size below what the formula suggests.
FAQ
Is optimal f the same as Kelly? They target the same thing but from different inputs — Kelly from odds, optimal f from your actual largest loss. On a real trade record with an outsized worst loss, optimal f is usually the more conservative figure. Compare with the Kelly criterion calculator.
What is f$? It is your largest loss divided by optimal f — the amount of account equity that should stand behind each single contract or unit. Bigger f$ means fewer, smaller positions.
How do I turn this into a stop and size? Once you know the equity behind each unit, pair it with the position size calculator and check the downside with the risk of ruin calculator before committing.