A stop-loss belongs at the level where your trade thesis breaks — not at a "comfortable" percentage from entry. First define the scenario and its invalidation point, then size the position from the distance to that point. A stop placed the other way around almost always sits in the crowd and gets swept before the move.
What it is — a pre-placed order that closes your position when price reaches the level where your trade scenario stops working. Not an insurance afterthought, but a formalized point of admitting you were wrong.
Why it matters — without a stop, the market decides the size of your loss. A stop turns undefined risk into a fixed number that Risk-Reward and position size are calculated from.
How to use it — invalidation level first, then distance, then size. Never in reverse. See how Moami's AI team checks your stop before you enter.
a proper stop locks in
must clear after the stop is set
the stop works while you sleep
position without a stop
What's inside
- What a stop-loss really is
- Stop types: market, limit, trailing, "mental"
- Where to place it: 4 steps from thesis to level
- Why "stop at −2%" doesn't work
- Volatility stops (ATR)
- Stop hunting: why everyone sees your level
- 5 signs your stop sits in the crowd
- Common mistakes
- Hard stop vs "I'll close it manually"
- Pre-placement checklist
- Frequently asked questions
What a stop-loss really is
Formally, a stop-loss is a conditional order the exchange executes automatically once price reaches your trigger. Practically, it is the only tool that moves your loss from the "whatever happens" category into the "exactly what I decided in advance" category. Before the stop is placed, your risk is a market variable. After — it's a constant in your plan.
The correct sequence of thinking goes: thesis first ("long, because the level was defended and bought up"), then the invalidation point ("the scenario is dead if price closes below the level"), then the distance from entry to that point, and only at the end — a position size that turns this distance into an acceptable risk per trade. In this chain the stop is not an order setting. It is a conclusion of your analysis.
The reverse sequence — "entered full size, now let me figure out where the stop goes" — produces stops anchored to comfort instead of market structure. Such a stop is almost always too tight, gets swept by ordinary noise, and the trade closes at a loss minutes before the move you were positioned for.
Stop types: market, limit, trailing, "mental"
Exchanges offer several execution mechanics, and the differences show up exactly when the stop matters most — during sharp moves.
| Type | How it works | The catch |
|---|---|---|
| Stop-market | On trigger, a market order is sent — the close is guaranteed | Slippage: in a thin market the fill can be noticeably worse than the trigger |
| Stop-limit | On trigger, a limit order is placed at your set price | A strong impulse blows through the limit — the position stays open |
| Trailing stop | The stop follows price at a fixed distance, locking in profit | A trail set too tight gets taken out by the first ordinary pullback |
| "Mental" stop | The level lives in your head; closing is manual | At night you sleep, during the day you hesitate. In practice it's no stop at all |
→ Scroll the table to the right
For most scenarios on crypto perpetuals the working choice is stop-market: a few tenths of a percent of slippage is almost always cheaper than an unfilled exit during an impulse. Stop-limit makes sense on liquid instruments in a calm regime, when you'd rather miss the fill than pay the slippage.
One more detail — the trigger source. Exchanges let you choose what counts as touching the level: last price, mark price or index price. For stops on perpetuals, mark price is the more reliable choice: it smooths manipulative last-trade spikes and is less likely to fire on a single anomalous candle inside a thin order book.
Where to place it: 4 steps from thesis to level
- Write the thesis down
"Long BTC: $76,500 has been bought up three times, structure is bullish." If the thesis doesn't fit in one sentence — there is no trade.
- Find the invalidation point
Where does the scenario stop working? Usually a close beyond a structural level: a swing low, a range boundary, an accumulation zone. Not "minus some percent" — a specific place on the chart.
- Add a noise buffer
Round numbers and obvious swings collect crowds of stops. Step beyond the level — past the wick of the previous sweep, below the round figure — so a noise spike doesn't kill a healthy trade.
- Size the position from the distance
The stop distance is now known — choose a size where getting stopped costs 0.5–2% of equity. If the structural stop demands too wide a distance, reduce size. Never tighten the stop instead.
Tip. After step 4, check the Risk-Reward ratio: if the distance from this stop to a realistic target works out worse than 1:2, the trade fails the filter — no matter how good the chart looks.
Why "stop at −2%" doesn't work
A fixed percentage is convenient for exactly one reason: it requires no thinking. But the market doesn't know where you entered, and it certainly doesn't know what percentage feels comfortable to you. BTC's volatility and a thin altcoin's volatility differ severalfold — the same "−2%" sits inside ordinary hourly noise for one asset and beyond the structural level for another.
A percentage stop fails in two symmetric ways. If the percent is tighter than the noise, the stop gets swept by a random wiggle — and a series of such sweeps breeds the durable belief that "stops don't work, I'll trade without them". If the percent is wider than the structure, you pay more than necessary for a broken thesis: the scenario died at the close below the level, but the position keeps bleeding until the arbitrary line is hit.
With leverage, percentage logic becomes doubly dangerous: liquidation distance shrinks, and a "comfortable −5%" at 20× means the exchange liquidates you before your stop ever triggers. The stop must always sit closer to entry than the liquidation price — otherwise its job is done by the exchange, at a worse price, plus the forced-close fee.
Volatility stops: the middle ground between percent and structure
Between a "naked percentage" and pure structure there is a working compromise — the volatility stop. The distance is tied to the instrument's average range, ATR (Average True Range), over a chosen period. The stop goes, say, 1.5–2 ATR from entry: for a calm BTC that's one distance, for a runaway altcoin — a very different one, and both automatically reflect the current market regime rather than your mood.
A volatility stop doesn't replace structure — it insures against misreading it. If 2 ATR lands closer than the structural level, the market is too noisy for this trade, and the honest conclusion is to skip it, not to pick "whichever stop looks nicer". When the structural level and the volatility distance coincide — that, on the contrary, is the mark of a quality entry.
Stop hunting: why everyone sees your level
Clusters of stops are liquidity. A large participant who needs to build a position needs counterparties, and a dense pack of stops under an obvious level is the most predictable place to find them. Hence the recurring pattern: price pokes below a swing low, collects the stops and liquidations, and reverses in the original direction.
This isn't a conspiracy — it's mechanics. Most traders' stops sit in the same places: under round numbers, under the last low, at textbook levels. The liquidation map shows these concentrations literally — as bright clusters that price visits on schedule. Your job is not to park your stop where the whole exchange can see it, but to hide it behind zones whose breach genuinely breaks the structure.
An extra filter is positioning skew. When the long/short ratio shows a crowd on one side, that crowd's stop clusters become the priority target: the market sweeps them first, and only then does the move begin. If your stop sits with the majority's stops — you are first in line.
5 signs your stop sits in the crowd
- Right under a round number. $80,000, $3,000, $100 — the levels everyone watches, and the stops beneath them get taken first.
- Tick-perfect under a swing low. The last low is the most obvious point on the chart. A stop at the low is a stop in the epicenter of the crowd.
- On a bright liquidation-map cluster. If your stop coincides with a dense liquidation band — that is exactly where price will come.
- Distance tighter than the instrument's typical noise. A stop closer than the usual one-to-two-hour range isn't protection — it's a lottery on the next candle.
- The stop has already moved since entry. If the level has "relocated" away from price, it's anchored to emotion, not structure — and almost certainly sits in a random spot.
Important. Getting swept before a reversal is not a reason to abandon stops. It is a reason to place them beyond structural invalidation zones, with a buffer for the wick, and to check the level against the liquidation map before entering.
Check your stop before the market does
Paste a planned or open position. Moami's AI team checks your stop against market structure, the liquidation map and your liquidation distance — and returns an explanation of where the scenario actually breaks, in about thirty seconds.
Common mistakes
Moving the stop as price approaches. One relocation turns a planned 1% risk into an unplanned 3–5%. A stop that moves is not a stop — it's a decoration.
Removing the stop "because of the news". The fastest moves happen exactly then. Dropping your protection at the moment of peak volatility is the worst possible timing.
Setting the stop and the leverage independently. Stop from structure first, then a leverage at which liquidation sits beyond the stop. The reverse order routinely ends in liquidation before the stop fires.
Re-entering immediately after a stop-out. That's not a plan — that's revenge. A new entry requires a new thesis, written from scratch.
Hard stop vs "I'll close it manually"
- Works 24/7 — crypto doesn't wait for you to wake up
- Executes without emotion at the most emotional moment
- Makes risk per trade measurable before you enter
- Keeps your statistics honest: stop fired — thesis was wrong
- Insures against cascade scenarios: gaps, impulses, liquidation cascades
- Requires being at the terminal during the move — i.e., luck
- In drawdown the brain searches for reasons to wait, not to act
- Risk per trade stays unknown until the very end
- On leverage, "didn't make it in time" ends in liquidation
Editorial note. Moami AI explains market mechanics and supports discipline — it does not give trading advice and does not guarantee outcomes. Any leveraged trade can end in the loss of the entire margin. This is educational material, not financial advice.
Pre-placement checklist
- The thesis is written in one sentence and names the invalidation level
- The stop sits beyond structure with a wick buffer — not tick-perfect under the low
- The level is checked against the liquidation map: the stop is not inside a bright cluster
- Position size is derived from the distance: a stop-out costs 0.5–2% of equity
- Liquidation price sits farther than the stop — the exchange won't close you first
- R:R from stop to target clears 1:2, calculated without rounding in your favor
A stop-loss is the cheapest risk-management tool in existence: it's free, automatic and works around the clock. Only one thing makes it expensive — the habit of placing it where it feels comfortable instead of where the thesis breaks. If this article makes you re-check even one of your stops against structure and the liquidation map, it has already paid for itself.