Where to put the stop loss — and what it costs your win rate

The single most underrated decision in any trade is where the stop goes. Most retail traders treat it like a footnote — "I'll put it just below support" — without realising that the placement of that line is the largest input to their win rate. Move the stop further from entry and the win rate goes up. Move it closer and the win rate goes down. The relationship is mechanical and roughly independent of how good the analysis was.

This isn't a soft observation about discipline. It's geometry. Price has to travel further to hit a wider stop, so wider stops get hit less often. Price has to travel further to reach a wider target, so wider targets get reached less often. The two effects pull in opposite directions, and where you place the stop relative to the target sets the equilibrium. A 70% win-rate strategy with 1:1 R:R and a 50% win-rate strategy with 2:1 R:R can have identical expectancy — the only difference is the stop. They feel like completely different strategies. They're the same strategy with the stop in different places.

This post walks the math, the four common stop-placement methods (fixed distance, ATR-based, structure-based, volatility-targeted), the trade-offs each one makes between win rate and average R, why "stop hunting" exists at obvious levels, and how to pick a stop that matches your strategy's actual edge instead of one that just feels comfortable. The interactive at the end lets you slide the stop distance and watch win rate, average R, and net expectancy move together.

The geometry of touch probability

Imagine a price that walks roughly randomly with a small drift in your favour — that's the rough shape of any market your strategy thinks it can predict. The probability that price touches a level X% away from entry before reaching some target Y% in the other direction is, for a near-symmetric process, approximately Y/(X+Y).

Plug numbers in.

• Stop 1% away, target 1% away: P(reach target) = 1/2 = 50% win rate.
• Stop 2% away, target 1% away: P(reach target) = 1/3 ≈ 67% win rate.
• Stop 3% away, target 1% away: P(reach target) = 1/4 = 75% win rate.
• Stop 0.5% away, target 1% away: P(reach target) = 1/1.5 ≈ 33% win rate.
• Stop 0.5% away, target 2% away: P(reach target) = 0.5/2.5 = 20% win rate.

That's it. That's the entire mechanics. The further away the stop, the higher the win rate. The further away the target, the lower the win rate. Combine them and the ratio determines where you land. A trader who wants 80% wins is implicitly saying "my stop is at least 4× further than my target." A trader who wants 3R wins on average has signed up for at most 25% wins.

The math here is the Brownian-motion exit-probability formula applied to retail trading. Real markets aren't pure Brownian motion — they have drift, fat tails, and microstructure effects — but for short-horizon retail trades the approximation holds remarkably well. If you back-test a non-trending random-walk model against real BTC perp data over a few hours, the touch probabilities for symmetric stop/target distances are within a few percentage points of 50/50.

The implication: if your strategy isn't doing better than this baseline, your win rate is coming entirely from stop placement, not from edge. A "70% win rate" with a 3:1 stop:target ratio is what a coin flip would produce with that ratio. If you're not beating the geometric expectation, the strategy has no edge — it just has a comfortable shape.

Four ways to place a stop

The placement methods retail traders actually use, in order of popularity:

1. Fixed-distance stops. "I'll always use a 1% stop." Simplest, easiest to explain, worst at adapting. A 1% stop on BTC during quiet hours is generous; the same 1% during a news release will be hit by noise alone. If your strategy doesn't account for changing volatility, your fixed-distance stop is doing the work of detecting volatility regimes for you, badly.

2. ATR-based stops. "I'll use 2× ATR(14)." Average True Range was published by Welles Wilder in his 1978 book New Concepts in Technical Trading Systems as a way to measure recent volatility. A 2× ATR stop scales automatically — when volatility expands, the stop widens; when it contracts, the stop tightens. The math behind why this works: if recent moves are typically 1×ATR and you place the stop at 2×ATR, you're requiring an above-average move to stop you out, which keeps the noise-driven hit rate low while preserving the ability to stop on real adverse moves.

3. Structure-based stops. "I'll put the stop just below the recent swing low." The most common method among discretionary traders. The logic: if price breaks a meaningful structural level, the trade thesis is invalidated, so the stop should be just on the other side of that level. The drawback: those levels are obvious to everyone, including the people whose business is sweeping liquidity from retail orders parked in the same place.

4. Volatility-targeted stops. "My stop should be calibrated so that the dollar risk per trade equals 1% of my account." This shifts the question from "where should the stop be?" to "given where the stop has to be for the trade to make sense, how big should the position be?" The stop placement gets dictated by the chart structure; the size flexes to keep the dollar risk constant. This is the method most professional risk managers use.

Each method picks a different point on the win-rate-vs-R:R curve. None is inherently better — they fit different strategies and different trader temperaments. But the choice is consequential and most traders make it without thinking about which trade-off they're accepting.

The win-rate / R:R seesaw

Two curves moving in opposite directions, crossing at 1:1 — the symmetric stop. Move left of centre and you've got a wide-target / tight-stop strategy: low win rate, big winners. Move right and it's tight-target / wide-stop: high win rate, small winners on average. The total expected value, before any actual edge, is the same at every point on the line. What changes is the shape of the equity curve and the psychological cost of trading it.

Pick your shape

The default starts at a symmetric 1%/1% with a 5-point edge above the coin-flip baseline and 0.12% round-trip cost — roughly the cost floor for retail BTC perps from the costs post. Move the sliders and watch the per-trade expectancy stay roughly constant unless you change the edge — confirming that stop placement alone doesn't create profit, it only redistributes shape.

The most useful experiment: leave edge at +5 and slide stop and target around. The expectancy stays positive across most settings. Then drop edge to 0 (no edge above the geometric baseline) and watch every configuration go negative once costs are subtracted. Without an edge, no stop placement saves you — you're just choosing what colour the equity curve will be while it grinds toward zero.

Why obvious stops get hunted

If you place your stop where every other retail trader places theirs — just below the recent swing low, or at the round-number support — you're parked in a queue with thousands of other resting orders. That's a target.

Aggregate stop-loss orders create liquidity pools visible to anyone watching the order book or the open-interest data. A market participant looking to enter a position can deliberately push price into one of these pools, trigger the cascade of stops, and absorb the resulting market orders at a price advantageous to them. This is referred to colloquially as "stop hunting" and is well-documented in the academic literature on liquidity events — see for example Kissell's The Science of Algorithmic Trading and Portfolio Management for the institutional view, and Lo & MacKinlay's research on the predictability of price movements around technical levels.

The defence isn't to remove your stop. It's to place it somewhere less obvious. Two common practical choices:

Beyond the level, by a meaningful margin. Instead of "just below" the swing low, place the stop a sensible distance further. The trade-off: a wider stop means smaller position size for the same dollar risk, plus the higher win rate / smaller average R shape from the geometry section. Often worth it.

Use a structure that isn't on a chart. Use a stop based on something the order book doesn't see — your strategy's invalidation logic, a time-based exit, a volatility-spike trigger. These are harder to harvest because they don't sit at fixed price points.

The third option, much-debated, is to use a mental stop — track the level and exit manually if it's hit. This works in theory and fails in practice because most retail traders, when watching their stop level get attacked, will move it rather than exit. The whole reason hard stops exist is to remove the decision from the moment of pressure. Mental stops give it back.

How to actually choose

For a strategy you're newly building or evaluating, a useful sequence:

1. Pick the stop method that matches your strategy timeframe. Scalping needs ATR-based stops because volatility shifts quickly. Day trading benefits from structure-based stops because charts have visible levels. Swing trading often does best with volatility-targeted stops because position size has to flex with multi-day moves.

2. Once the stop method is fixed, the target is a free parameter. This is where the win-rate-vs-R:R trade-off lives. Most retail strategies do best at 1.5:1 to 2:1 reward-to-risk, which puts the win rate around 40-55%. High-win-rate strategies above 70% are achievable but require small targets that get eaten by costs faster.

3. Don't move the stop after entry. "Trailing stops" are seductive and almost always reduce expectancy. The math: trailing locks in winners but converts what would have been larger wins into break-even outcomes when noise pulls back. Run the back-test of your strategy with and without a trailing rule and the version without usually has higher net expectancy. If the without-trailing version has unacceptable drawdown shape, the answer is smaller size, not trailing.

4. Stop hunting is real but fixable. Place stops sensibly beyond obvious levels. Don't park orders right where everyone else does. Accept that liquidity pools exist and design around them.

For the broader picture of how stop placement interacts with the strategy's win rate as a number you can actually trust, the win-rate-sample-size post covers why the win rate from your first 50 trades doesn't yet pin down which point on the seesaw you're actually at. And the edge-after-cost post covers why the cost floor pushes you away from very tight stops in the first place — small targets get eaten alive by commissions and slippage. Once you've decided where the stop goes, the position size calculator turns that distance into the exact coin amount to take so the stop fires for the percentage of your account you actually meant to risk.