How does fixed fractional sizing work?
Fixed fractional sizing is the rule that risks a constant fraction of current equity on every trade, rather than a constant number of dollars or a constant number of shares. It is the most common sizing framework in the trading literature, and its behaviour under losing streaks is the reason.
The rule in one line
Risk on the next trade = current equity × f, where f is a fixed fraction.
Because f multiplies current equity, the dollar bet shrinks automatically after losses and grows automatically after wins. Nobody has to decide to "trade smaller after a bad month" — the arithmetic already did.
Worked: ten losses in a row
Start with $100,000 and f = 1%.
- Trade 1 risks $1,000. It loses → equity $99,000.
- Trade 2 risks 1% of $99,000 = $990. It loses → $98,010.
- Trade 3 risks $980.10 → $97,029.90. And so on.
After ten straight losses the equity is 100,000 × 0.99¹⁰ = $90,438. Total damage: 9.56%, not 10%. The tenth trade only risked $913.52, because by then there was less to risk.
Compare the fixed-dollar alternative — risk a flat $1,000 every time. Ten losses removes exactly $10,000, leaving $90,000. Slightly worse in absolute terms, and worse in a subtler way: that tenth $1,000 bet was 1.10% of the $91,000 that remained. The fixed-dollar rule quietly increases the fraction at risk exactly when the account can least afford it. Fixed fractional does the opposite.
What the fraction costs you
Doubling f slightly under-doubles the drawdown — compounding on a shrinking base sees to that, since each loss is taken on what the last one left. What more than doubles is the climb back, because D ÷ (1 − D) is convex:
f |
Equity after 10 losses | Drawdown | Gain needed to recover |
|---|---|---|---|
| 1% | $90,438 | 9.56% | 10.6% |
| 2% | $81,707 | 18.29% | 22.4% |
| 5% | $59,874 | 40.13% | 67.0% |
You will see "1–2% per trade" repeated constantly in trading books and courses. Treat it here as a convention people quote, not an instruction — this course's job is the table above, which shows what the convention actually buys. What fraction is appropriate for any individual depends on things no lesson can know.
The catch
Fixed fractional protects the account arithmetically, not physically. It assumes the loss is really f — which assumes the exit happens where you planned. Unit 2 is about the several ways that assumption breaks.
It also has a mathematical quirk worth knowing: because each trade multiplies equity by a factor, the account can theoretically approach zero without ever reaching it, while in practice minimum position sizes, fees and margin rules put a hard floor under "too small to trade."
Try it now
- Start from a hypothetical $60,000 and apply eight consecutive losses at
f = 1.5%(multiply by 0.985 eight times). What is the ending equity and the drawdown? - Repeat with a fixed-dollar rule of $900 per trade. Which ends higher, and what fraction of equity was the eighth bet under each rule?
- Now run five wins of 2× the risk through the same rule from $60,000 (each trade risks 1.5% and gains 3%). Note that the mechanism which softens streaks also compounds them.
- Ground the abstraction: take the latest close from the month below and work out how many shares $900 of risk actually buys at a $2.00 stop distance. Drop a Level $2.00 under it if you want to see how ordinary a distance that is. The fraction is a concept; the share count is the thing you would have to type.