Position sizing is the only edge you fully control
Most traders spend years hunting for a better entry. The variable that actually decides whether an account survives a bad month sits somewhere else entirely.
Ask a room of traders what they are working on and almost everyone will describe an entry problem. A cleaner signal, one more confirmation, another filter to cut false breakouts. Very few will say they are working on how much they risk per trade — and that is the one input entirely theirs to set.
Entries live in the market's hands. You can improve them, but you cannot control what price does after you click. Position size lives entirely in yours. It is decided before the trade, it is arithmetic rather than judgement, and it is the difference between a losing streak that stings and one that ends the account.
The maths that changes behaviour
Recovery from a drawdown is not symmetrical, and the asymmetry gets brutal faster than intuition suggests.
| Drawdown | Gain needed to recover |
|---|---|
| 10% | 11.1% |
| 25% | 33.3% |
| 50% | 100% |
| 75% | 300% |
| 90% | 900% |
At 10% down you need a decent month. At 50% down you need to double the account just to get back to flat — with the same system that just lost half of it. Most traders do not accept that arithmetic. They increase size to catch up, and the curve accelerates downward.
Risk of ruin is a real number
Take a system that wins 45% of the time at 2R per win. Positive expectancy — a good system. Now run it at different risk levels and the picture changes completely.
| Risk per trade | Worst realistic drawdown | Practical outcome |
|---|---|---|
| 0.5% | ~8% | Uncomfortable, recoverable |
| 1% | ~16% | Manageable |
| 3% | ~45% | Most traders quit here |
| 5% | ~65% | Effectively over |
The system did not change. Its edge did not change. Only the size did — and only the 0.5% and 1% versions are ones a human can actually sit through. A losing streak of eight or nine trades is completely normal at a 45% win rate. It will happen to you, probably this year.
Risk small enough that being wrong five times in a row is boring, not fatal.
What a rule actually looks like
A workable rule is specific, written before the session, and does not change because a setup looks unusually good. Ours has three parts.
1. Fixed fractional risk
A set percentage of current equity per trade, recalculated as the balance moves. Losing shrinks your size automatically; winning grows it. This is the single most important property — it makes the account self-correcting.
2. Stop-defined sizing
The stop goes where the idea is invalidated. Lot size is then derived from that distance. Never the reverse. Choosing size first and then placing the stop where it fits is how traders end up with stops parked inside noise.
The calculation for a 1% risk on a $10,000 account with a 40-pip stop on EURUSD:
| Input | Value |
|---|---|
| Equity | $10,000 |
| Risk % | 1% = $100 |
| Stop distance | 40 pips |
| Value per pip needed | $100 / 40 = $2.50 |
| Position size | 0.25 standard lots |
3. A daily and weekly cap
Two or three losses in a day and the platform closes. Not a suggestion — a hard stop.
The part people skip
The third rule is the one abandoned first, and it is doing most of the work. A per-trade limit protects you from a bad trade. A daily limit protects you from a bad state — the hour after a painful loss, when judgement is measurably worse and the urge to make it back is strongest. Almost every account-ending day in trading history was a series of revenge trades, not one bad position.
Where correlation quietly breaks the rule
Three positions at 1% each is not 3% of risk if those positions are correlated. Long XAUUSD, short USDCHF and long EURUSD during a dollar move is effectively one trade in three costumes. If the dollar goes the wrong way, all three lose together.
Practical fix: cap total open risk, not just per-trade risk. Two or three percent across everything open, with correlated positions counted as one.
Where to start this week
Take your last fifty trades and recalculate them at a fixed fractional risk. Not to prove the system was profitable — to see how the equity curve's shape changes once size stops varying with conviction. For most traders the peaks get lower and the valleys get much shallower. The shallower valleys are what keep you in the seat long enough for an edge to show up.
If you are automating this, the risk module belongs in the code before any entry logic — see building your first EA: structure before strategy.
This article explains a risk framework used in our teaching. It is not investment advice, a recommendation, or a guarantee of any outcome. Every trader's capital, risk tolerance and circumstances differ — decide your own rules, and consider speaking to a registered adviser before committing capital.