How to Actually Set a Stop Loss: Three Methods and One Hard Rule

When most people talk about stops they’re really talking about “how much loss I can tolerate emotionally.” That’s emotional pain, not a trading rule. A real stop is a pre-written order: triggered, executed, no analysis, no “one more candle.” This piece skips the motivational tone and lays out three working methods, where each fits, and the details that quietly wreck accounts.

The hard rule that comes first

Every stop-loss method shares one prerequisite: you write the level down before you enter the trade.

Note app, sticky paper, the actual stop order — anywhere. But never after the position is already underwater. The reason is unforgiving: a brain looking at an open loss automatically inflates the “it’ll come back” probability and shrinks the “it might keep falling” probability. You think you’re analyzing; you’re actually consoling. Every blow-up story — including system-wide cascades like the Luna and UST collapse — carries the same hidden line: “I thought it would come back.”

Method 1: Structure stop — let the chart speak

The structure stop is the closest thing to “listening to the market.” Find the most meaningful nearby support or resistance, then put the stop on the other side of it.

Examples:

• Long: find the most recent clear swing low A, place the stop 0.5%–1% below A.
• Short: find the most recent failed swing high B, place the stop 0.5%–1% above B.

Why the buffer? Market makers and large players know retail piles orders at round numbers and obvious swing points. A stop sitting exactly on A gets wicked out very often. A little extra room is the tax you pay to noise.

When it works: trending or breakout conditions with clear structure. Range markets puncture structure repeatedly and structure stops misfire.

Method 2: Percent stop — simple, but only with position sizing

A percent stop means exit when you’re down X% — 2%, 5%, 10%. The strength is that it’s mechanical and judgment-free. The weakness is that it ignores volatility.

A 5% stop on BTC may be a routine retrace; on a high-vol altcoin it’s almost “stopped at entry.” So a percent stop can’t run alone — it has to be tied to position size:

• high-vol asset → smaller size, wider percent stop;
• low-vol asset → larger size, tighter percent stop.

The mechanics of working backward from “max loss per trade” to a concrete position size are covered in position sizing and per-trade loss limits. Here the point is just this: you’re not picking a comfortable percentage, you’re picking how much of the account you’re willing to lose in one trade. The second decision has to exist before the first one makes sense.

Method 3: Volatility stop (ATR) — different space for different assets

The volatility stop uses ATR (Average True Range) — roughly, “how much does this asset typically move per period?” Any chart tool computes it for you; you just use it.

The common form:

• long stop = entry − K × ATR;
• short stop = entry + K × ATR.

K is usually 1.5, 2, or 2.5. Higher K, more noise tolerance, larger loss when wrong. Lower K, more sensitivity, higher odds of being shaken out. A common pairing: K=2 on 4h, K=1.5 on 1h, K=3 on daily.

The big benefit: the stop adapts to the asset. BTC in a quiet stretch has low ATR and a tight stop; an altcoin in a flush has high ATR and a wider stop. It fixes the “one number for everything” flaw of plain percent stops.

Three methods side by side

My own use is structure stop as the primary, ATR as a sanity check: draw the structural level first, then look at how it sits in ATR terms. If the structural stop is implausibly close, the position size has to shrink; if it’s very far, the size has to shrink more.

Three details that quietly wreck people

First, don’t replace a real stop order with a “mental stop.” You believe you’ll click sell when price gets there. You won’t. The moment it arrives, you start hunting for reasons it might bounce. Place the limit or market stop in advance and let the rule execute outside of you.

Second, never move a stop in the losing direction. This is where every blown account starts. Permitting yourself to move it “just once” is the same as throwing away the analysis you did before entering. Move stops only in the profit direction (a trailing stop). Never against you.

Third, a stop getting hit is not failure — it’s the rule working. People wreck themselves by getting stopped a few times in a row, deciding stops are the problem, and removing them. Then one large move clears the account. A few stop-outs are simply the cost of probability. As long as the loss per trade stays inside your tolerance, the long run is positive. That’s why “max loss per trade” must be decided before you trade — it determines how many shakeouts you can absorb. The fuller view is in crypto risk management.

The crypto-specific case: wicks and extreme moves

Crypto wicks can be absurd. A 5% flash drop on an altcoin is a routine Tuesday. Two ways to handle it:

• Use a limit stop, not a market stop: limits cap your fill price but may not execute in extreme moves, so you need to be watching.
• Treat wick losses as the entry fee for high-vol markets: smaller size, wider stop, which paradoxically is steadier overall.

Both beat “watch the wick, hold and hope.” When the cascade actually arrives, big money moves before you do; what holding through it looks like is documented in lessons from the FTX collapse — plenty of people held through it, very few of them are still in the game.

Closing

A stop loss is an admission that you can be wrong. Its purpose isn’t to predict the market; it’s to lock the maximum loss in a cage before you’re wrong. You don’t need to be right every time, you only need to not die when you’re wrong. Put that mindset above all three methods and any of them works.