The Two Jobs of a Protective Stop
Protective stops are often described with a simple image: they work like a red light. A driver can run a red light, but doing so is never truly safe. In trading, a stop is the price level at which a position is closed to limit damage, and the study notes behind this article treat that level as a structural part of a trading plan rather than a matter of personal comfort. Two functions stand out. First, the stop defines the maximum loss a single position can realistically incur, normally written as R — the initial risk per trade. Second, it creates a benchmark: once R is set, every later gain can be expressed as a multiple of that initial risk, which is what makes outcomes comparable across trades and across systems. Without a defined R, a profit of $1,500 has no context; with R defined, the same number becomes 20R, 9R, or something else depending on how much was risked. The notes frame the trader's central task as designing a profit plan capable of returning several times R. The stop does not create an edge by itself; it defines what is at stake while the market decides whether the idea works. A second theme is noise. A stop placed too close to the entry can be triggered by ordinary price fluctuation rather than by a genuine change in the market's behavior. To move the stop beyond that noise, the notes suggest three routes: set it at a multiple of ATR (average true range), use a Dev-stop based on the statistical distribution of price movement, or estimate the maximum adverse excursion (MAE) of comparable trades and place the stop beyond that figure. Finally, stops are described as worst-case protection, not as guarantees. In fast or gapping markets, the executed price can be worse than the level specified, so the maximum loss is a planning assumption rather than a certainty.
William O'Neil's CAN SLIM: The 7–8% Loss Rule
William O'Neil's CAN SLIM method is unusual among stock systems in that it deliberately avoids stops tied to overall market conditions. Even so, O'Neil is explicit that a stock should never be allowed to move 7% to 8% or more against the entry. This is a percentage giveback measured from the purchase price, not a percentage of account capital. The distinction matters. If a stock is bought at $20, the tolerated adverse move is roughly $1.40 to $1.60. If a stock is bought at $100, the tolerated adverse move is $7 to $8. The rule scales with the price of the instrument rather than with the size of the account, so two positions of equal dollar value but different share prices will have different dollar risk. O'Neil also warns that 7% to 8% represents the maximum tolerable loss, not a target. In practice, he suggests the average loss across all trades should sit closer to 5% to 6%, which implies that many positions are cut before reaching the outer limit. Entry timing supports this logic, because CAN SLIM favors buying as a stock emerges from a base or consolidating pattern, when a well-chosen entry should not immediately travel far against the buyer. The study notes describe O'Neil's guideline as one of the better stock-trading frameworks available, while arguing that it can still be improved. The suggested improvement is to use market-based stops instead of a fixed percentage, since a fixed 7% to 8% applies the same dollar distance to a slow, low-volatility stock and to a fast, high-volatility one. A volatility-aware stop, by contrast, adapts to the instrument being traded.
