The standard sizing rule says risk a fixed fraction of the account, commonly one percent, on any single trade. Size is then set so that the distance to your stop equals that fraction.
The rule is sound and it is incomplete, because it accounts for the loss if your stop is hit and ignores the cost you pay whether it is hit or not.
The stop is not the only thing you pay
If you risk one percent to your stop and your round trip costs the equivalent of a fifth of that distance, then a stopped-out trade loses one percent plus the toll, and a trade that goes nowhere still loses the toll.
Over a long sequence, the trades that went nowhere are usually the most numerous. They do not appear in a win rate and they do not appear in a risk-per-trade calculation, but they are the quiet drain on the account.
Why tight stops are disproportionately expensive
Cost is charged on notional, not on the distance to your stop. Halving your stop distance lets you double your size for the same one percent risk, which doubles the notional, which doubles the cost you pay in currency terms.
So a tighter stop does not merely increase the chance of being stopped out by noise. It increases the toll on every trade, at the same time. The two effects push the same way, and this is the mechanism behind the common experience of a strategy that got worse after its stops were tightened.
A worked comparison
- Account 10,000. Risk one percent, so 100 per trade.
- Wide stop, 2 percent away: position size 5,000. At an 11 bps round trip, cost is about 5.50.
- Tight stop, 0.5 percent away: position size 20,000. Same 11 bps round trip, cost is about 22.00.
- Same one percent of risk, same rule followed correctly, four times the cost per trade.
Where the spread fits
The spread also interacts with stop placement directly. A stop placed inside the typical spread is not a risk level, it is a guarantee of being taken out by ordinary quoting. In thin markets the spread alone can exceed a stop distance that looked reasonable on a chart drawn from mid prices.
Charts are usually drawn from one price per period. Your fills are not. Any stop distance close to the spread should be checked against the actual book rather than the line on the chart.