A stop-loss strategy is the pre-committed rule that defines where, and on what basis, you will exit a losing (or maturing) position before discretion and hope take over. It is the operational heart of risk management: the stop distance, not the entry, is what fixes how much a single trade can cost, and through position sizing it determines how large that trade may be in the first place. This section catalogues the recognized methods of stop placement — how the exit level is derived — which fall into two orthogonal families: rules that bound the magnitude of a loss (percentage, volatility, structure, trailing) and a rule that bounds its duration (time stops). The core tension running through all of them is the same trade-off in different clothing: a stop placed tight keeps each loss small but ejects you on ordinary noise; placed loose it survives noise but makes each loss larger. There is no setting that escapes this, and — as the academic evidence below shows — there is no method that is universally "best." The right choice is conditional on the strategy, the instrument, the timeframe, and the volatility regime.

What the section covers — and what it defers

These nodes describe how to derive the exit price. They deliberately do not cover the downstream step of converting that exit distance into a share count — that is the Position Sizing family (fixed-fractional / percent-risk sizing, volatility-based sizing), which consumes the stop distance produced here. They also defer the chart definitions of the anchors a stop hangs on (swing lows, support/resistance, breakout pivots, moving averages) to the Technical Analysis branch, and the exact swing entry/stop/target/hold mechanics to the Swing Trading branch. The relationship to remember: the stop method here defines "1R" of risk, and everything downstream is built on that unit.

The sub-topics (the methods)

• Fixed-Percentage Stops — the simplest rule: exit at a constant percentage below entry (O'Neil's 7–8% CAN SLIM cut being the canonical instance). Mechanically un-rationalizable and easy to automate, but blind to how much a given stock "breathes." Best understood as a behavioral discipline tool and an outer sanity-ceiling, not the primary stop on a volatile name.

• Volatility (ATR) Stops — places the stop a multiple of Average True Range (Wilder, 1978; typically 1.5–3×, a convention not an optimum) from the reference price, so the leash widens for jumpy stocks and tightens for quiet ones. Removes the arbitrariness of the distance but not the judgment of how much room is enough. The Chandelier Exit is its canonical trailing form.

• Structure-Based Stops — the mainstream discretionary default: place the stop just beyond the price level whose violation would invalidate the trade thesis (swing low, pattern low, support, reclaimed pivot). It answers "at what price am I wrong?" rather than "how much will I lose?" — and lets the resulting distance dictate size. Its weakness is subjectivity and stop-clustering at obvious levels.

• Time Stops — the orthogonal axis: exit after N bars/days regardless of price. It controls loss duration, not magnitude — an opportunity-cost tool, not a loss-control one. Most defensible where the edge has a built-in clock (mean reversion, event trades); generally inappropriate for trend/momentum setups, where it truncates the winners that pay for everything.

• Trailing Stops — the exit engine for "let winners run": a stop that ratchets in your favor and never moves backward, converting open profit into a locked floor. Implemented via fixed %, ATR/Chandelier, structure (higher swing lows), or a moving-average cross. The defining trade-off — tight whipsaws, loose gives back — is unavoidable.

When stop methods matter most — and least

Stop choice matters most where price action contains real noise relative to the trade's edge: volatile, gappy, or thinly traded names, leveraged positions, and tight-entry breakout methods where a small adverse move genuinely invalidates the setup. It matters least — and can actively harm — in mean-reversion strategies whose entire thesis is buying the dip you'd otherwise be stopped out of. The method also interacts with structure: a clean ATR or percentage distance that slices through major support buys you nothing a structure-aware stop wouldn't have respected, which is why practitioners frequently combine methods (anchor at structure, buffer with ATR).

Adoption, debate & evidence

Stop-loss discipline is near-universal as taught practice across retail, prop, and discretionary trading. But the claim that stops improve returns is genuinely conditional, and the section's children are careful to say so. The anchor result is Kaminski & Lo, When Do Stop-Loss Rules Stop Losses? (Journal of Financial Markets, 2013/2014): under a random-walk price process a stop-loss rule always reduces expected return (you pay the round-trip cost for no informational benefit); stops generate a positive "stopping premium" only under momentum / regime-switching dynamics, and they can destroy value for mean-reverting strategies by exiting right before the reversal. Their US-equity test (1950–2004), with stop-out proceeds parked in long-term bonds, found roughly 50–100 basis points per month added during stop-out periods — but the authors note this is partly a flight-to-bonds effect, not proof the threshold itself is optimal. The honest summary across methods: stops reliably truncate the left tail and enforce discipline, but they are not a standalone source of edge, and their net return effect depends on regime, asset, costs, and setting. Specific marketing figures (e.g. "ATR stops gave 24% higher returns," "best N-day exit is X") almost always trace to single un-peer-reviewed backtests and should be treated as illustrative, not established — several are flagged as disputed in the child nodes.

Strengths & limitations (at the section level)

Strength: every method here, applied consistently, guarantees no single trade becomes catastrophic and removes the in-trade discretion that is the dominant failure point for novices. Limitation: no method dissolves the tight-vs-loose trade-off, and each can be the wrong tool — a flat percentage on a high-beta name, a hard time stop on a trend, a trailing stop on a mean-reversion-to-target setup. The single most common misuse across the whole family is moving the stop to fit a desired position size (tightening it to buy more shares) instead of moving the size to fit a logically placed stop — which silently converts a risk-control tool into leverage and produces a string of noise-driven losses.

Sources

• Kaminski, K. & Lo, A. W., When Do Stop-Loss Rules Stop Losses? — Journal of Financial Markets (2013/2014); SSRN 968338; MIT DSpace. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=968338 (random-walk reduces return; momentum/regime-switching positive premium; mean-reversion value destruction; 50–100 bps/month during stop-out periods)
• StockCharts ChartSchool — Average True Range (ATR), Chandelier Exit (Wilder 1978; Le Beau; 22-period / 3× defaults).
• O'Neil, W. J., How to Make Money in Stocks — 7–8% cut-loss rule.
• Cesar Alvarez, N-Day Exits with Mean Reversion — weak/inconsistent time-stop evidence.
• López de Prado, Advances in Financial Machine Learning (2018) — triple-barrier (vertical time barrier).
• Child nodes in this section (fixed-percentage, ATR, structure, time, trailing) for method-specific mechanics and full source lists.

Disputed / flagged: "Stops improve returns" is conditional (Kaminski & Lo); precise method-vs-method outperformance figures and "best" day-counts in popular content come from non-peer-reviewed backtests and are illustrative only.