04 · Perpetual Trading in Practice & Risk Control

01-Perpetual Swaps explained "what the contract is"; this article explains "how to survive": how to pick leverage, how big a position to open, where to put the stop-loss, where the liquidation walls are, and how to exit after consecutive liquidations. Everything here is practical: every conclusion comes with a worked number, a checklist, or a template. All leverage tiers, rates, and liquidation data follow the latest rules of exchanges and data platforms.

⚠️ Risk Warning: nothing in this article is trading advice. The first step of perpetual trading is not "finding a way to make money" but "admitting there are ten thousand ways to lose it": fighting the trend, oversized positions, bag-holding, careless stops, wicks, funding, ADL... The checklists and templates below can only reduce mistakes, not eliminate risk. If you have already blown up 3+ times in a row, finish Section 7 before placing the next order.

---

1. Preparation: How to Choose Leverage

1.1 Leverage Is Not a "Multiple", It Is a "Distance to Liquidation"

Leverage basics (10x leverage dies on a 10% move) are covered in Core Trading Concepts. Recapping the conclusion of 01-Perpetual Swaps: an adverse move of roughly 1/leverage brings you to the edge of liquidation (after maintenance margin, fees, and funding, the actual forced liquidation triggers earlier).

Leverage × adverse move = the fraction of margin lost:

Adverse move	2x	5x	10x	20x	50x
2% against	−4%	−10%	−20%	−40%	Liquidation
5% against	−10%	−25%	−50%	Liquidation	Liquidation
10% against	−20%	−50%	Liquidation	Liquidation	Liquidation

The "Liquidation" cells are rough 1/leverage estimates; actual forced liquidation triggers earlier due to maintenance margin, fees, and funding (defer to the estimated liquidation price shown on the exchange).

Three conclusions:

1. 2x–5x is the "survivable" zone: there is still room to maneuver through ±10% BTC days, and the odds of riding out wicks are highest;
2. 10x is the per-trade ceiling for most veterans: liquidation at a 10% adverse move means stop-loss room of 3%~5% with about a full cushion left;
3. 20x and above only suits "very tight stop + very small size" quick trades: widen the stop a little and the position gets liquidated before the stop triggers — at that point the stop is decoration.

1.2 Why "Low Leverage + Big Position" Loses More Easily Than "High Leverage + Small Position"

Intuitively "low leverage = low risk", but that is only half true. Risk depends not on the leverage number but on the notional position (leverage × margin) relative to the account, and on whether the stop distance matches market volatility.

Worked example: account 10,000 U, per-trade risk budget 2% (200 U).

Plan	Leverage	Margin	Notional	Stop distance	Stop loss	Problem
A: low leverage, big position	5x	5,000 U	25,000 U	0.8%	200 U	A 0.8% stop is guaranteed to be swept by BTC's daily range
A': wider stop	5x	5,000 U	25,000 U	2%	500 U	One trade loses 5%, 2.5x over budget
B: high leverage, small position	10x	500 U	5,000 U	4%	200 U	Stop ≈ BTC's daily range; liquidation price ~9%, reasonable

• Plan A can only cap risk by squeezing the stop to 0.8% — a tight stop = high sweep rate; three sweeps in a day is −6% on the account;
• Plan B's notional is only half the account, so the stop can sit at 4% with the liquidation price (~9%) still far beyond it — the sweep probability is actually lower;
• Add slippage: Plan A's stop order is 25,000 U notional, so 1% slippage in a wick costs an extra 250 U; Plan B's stop order is only 5,000 U, far less hurt by the same wick.

Conclusion: "low leverage" is not a get-out-of-jail card — "low leverage + betting the whole capacity it frees up" is the most common way to die. Risk control only ever looks at two things: notional position / account equity, and whether the stop distance matches volatility. Get those right and the leverage number itself is irrelevant.

---

2. Position Management in Practice

2.1 Applying the 1%~2% Per-Trade Risk Rule to Perpetuals

The stock/futures rule "per-trade loss ≤ 1%~2% of the account" ports directly to perpetuals, with a single formula:

Position (notional) = Per-trade risk amount ÷ Stop distance
Required margin = Notional position ÷ Leverage

Worked example: account 1,000 U, stop 5%, leverage 10x

Item	Calculation	Result
Per-trade risk (1%)	1,000 × 1%	10 U
Notional position	10 ÷ 5%	200 U
Required margin	200 ÷ 10	20 U (2% of account)
Liquidation distance	~10% theoretical at 10x, ~9% after maintenance margin	Stop 5% sits inside it, with about a full cushion

If the same account opens 1,000 U notional (margin 100 U, 10% of the account): a 5% stop loses 50 U = 5% of the account. Four consecutive stops take the account from 1,000 U to 814 U — an 18.6% drawdown: the compounding damage of "lose a little each time, die in aggregate".

Two practical corollaries:

1. Position size is a function of stop distance, not of "feeling": the tighter the stop, the bigger the allowed position, but tight stops get swept more; the wider the stop, the smaller the position must be. When the two conflict, protect the stop room first;
2. Most people's sizing problem is really a "stop too tight" problem: rather than shrinking the stop to accommodate a big position, shrink the position to accommodate a sane stop.

2.2 Choosing Cross vs Isolated in Practice

01-Perpetual Swaps covered the mechanics: isolated uses only the position's own margin and liquidation loses only that position; cross uses the full balance, pushing the liquidation price farther but risking the whole account.

Scenario	Recommended mode	Why
Beginner / funds < 10k U / strict per-trade risk control	Isolated	The blast is contained in one position and cannot drag down the account
Large idle USDT balance acting as buffer	Cross	The balance automatically props up the liquidation price — free risk buffer
Hedged long/short (spot + contract, or two-way contracts)	Cross	Profits on one side replenish margin on the other, avoiding one-sided liquidation — but wrong direction kills both
Multiple same-direction high-leverage positions at once	Isolated	In cross mode, one blow-up can detonate the other profitable positions
Opening new positions while in drawdown	Isolated	Cross drags the new position into the old one's pit
Borrowed money	Neither allowed	Fix the source of funds before talking about trading

Practical advice: even in cross mode, size each trade's risk with an "isolated mindset". Cross is a tool to "push the liquidation price farther", not an excuse to bag-hold — bag-holders die in either mode; cross just makes the death more thorough.

---

3. Take-Profit and Stop-Loss in Practice

3.1 How to Set the Stop: Three Methods

Method	Rule	Strength	Weakness	Suits
ATR stop	Entry ∓ 1.5~2 × ATR(14)	Adapts to volatility; bigger swings = wider stop	Distance gets far in big moves, forcing smaller size	Swing, trend following
Structure stop	Below the key prior low / above the prior high, plus 0.5%~1% buffer	Grounded location; if swept, the reason is clear	Finding structure takes chart experience	Intraday, swing
Fixed-percentage stop	Entry ∓ 2%~5%	Simple, mechanical, executable	Disconnected from volatility; may be too near or too far	Beginners, systematic trading

Combined advice: anchor on structure, validate the distance with ATR, cap with a fixed percentage. When the three methods disagree, take the "farthest stop that still fits the per-trade risk budget", then back out the position (formula in 2.1).

3.2 The Relationship Between Stop Distance and Margin

The stop must sit inside the liquidation price, or the logic inverts: the position gets liquidated before the stop triggers, making the stop pointless. Since 1/leverage ≈ liquidation distance:

Fix the stop distance first → back out the leverage cap (leave one full cushion) → finally compute the position size

Stop distance	Theoretical leverage cap (1 ÷ stop distance)	Practical cap (leave ~2x cushion)
3%	33x	15x
5%	20x	10x
10%	10x	5x
20%	5x	2x

Check formula: stop distance × actual leverage < 1, ideally < 0.5. For example 10x with a 5% stop → 0.5, barely passing; 20x with an 8% stop → 1.6, the position always dies first.

3.3 The Reality of "Stops Get Swept by Wicks" and the Response

Reality: crypto stop orders are frequently swept by a wick after which price returns — "the direction was right after all". Two causes:

1. Stop orders are market orders: once triggered they fill at the going market price; when a wick slices through the stop level, the actual fill is far worse than the stop price (slippage);
2. Stop levels are highly predictable: exchange APIs, order-book patterns, and public liquidation data let programs spot them; dense stop clusters are liquidity "gold mines".

Responses:

Response	How
Buffer the location	Put the stop 0.5%~1% below the prior low, never "exactly at" it
Avoid round big numbers	Stops just under 30,000 / 60,000 / 100,000 are the most crowded
Avoid liquidation walls	See Section 4; stay away from clustered liquidation orders
Re-enter after a sweep	If the trend is intact, re-enter on the original signal; don't chase, don't revenge-add
De-leverage before big events	Cut size before CPI, rate decisions, ETF rulings, token unlocks so the liquidation price sits far from market — don't remove the stop, only reduce exposure

⚠️ Risk Warning: a swept stop is not the stop's fault. The stop's only job is "capping the per-trade loss", not "never getting hit". Removing the stop for fear of sweeps ("bag-holding") is the most classic blow-up path: one deep pullback without a stop eats all the profit saved by the previous 20 stops.

---

4. Liquidations and Liquidation Walls

4.1 How to Read Liquidation Data

The mainstream data platform is Coinglass (aggregating Binance, OKX, Bybit, and the whole market); exchange pages and market apps also ship "liquidation rankings / heatmaps" (data per the platform's latest statistics).

Data dimension	How to use it
By exchange / by coin / by direction	Judge which side — longs or shorts — carries more leverage in a coin and is easier to sweep
Aggregated by 1h / 4h / 24h	See the "tidal" rhythm of liquidations; avoid opening positions at liquidation peaks
Single liquidation leaderboard	Single liquidations of millions to tens of millions of USDT come from whales; dense zones sit near those price levels
Liquidation heatmap	Darker color = more pending liquidation orders clustered near that price, i.e. the "liquidation wall"

Note: liquidation data is statistics after forced liquidations have filled — lagging, not a real-time signal; it answers "where the wall is", not "whether the wall will be pushed".

4.2 "Both-Sides Blow-Up" Markets

The mechanism chain:

One side's high-leverage positions cluster (a liquidation wall) → price touches the wall → cascading forced liquidations (market orders)
→ The liquidation orders push price through the wall → the opposite side's positions also trigger → longs and shorts blow up in a chain
→ Derivatives prices deviate violently from spot → the mark price mechanism holds the line, but the spot index itself is also swinging violently

Classic case: May 19, 2021. BTC dropped from the $40k area to wick near $30k intraday (lower on some exchanges); per Coinglass and other platforms, total crypto contract liquidations that day ran to the billions of USD, with longs and shorts cascading — "both-sides blow-up" became a textbook day for the perpetual world.

A "double blow-up" is not an exchange malfunction but the mathematical inevitability of clustered high leverage + chained liquidation fills: with enough leveraged positions on both sides, sweeping one side necessarily hits the other. The lesson of 5·19: those without stops, without buffers, and oversized die on both sides in a double blow-up.

4.3 How Big Money "Hunts" Retail Stops

Hunting works because retail behavior is highly homogeneous: stops all sit "just below the prior low", "at round numbers", "right below structure levels", and these spots can be observed in advance via order-book patterns, cancel monitoring, and liquidation heatmaps.

The usual script:

1. Whales/algorithms lock onto the stops and liquidation orders clustered below a level;
2. Large market orders or consecutive sells smash through the level, triggering stop cascades and liquidation walls;
3. They fill in the deep spread created by slippage and forced liquidation (absorbing the swept liquidity);
4. Price reverts, and retail discovers "yet another wick".

Survival principles for the retail side:

• Distance your stop from the crowd: 0.5%~1% below the prior low; 1% below round numbers;
• Trim proactively before key levels: when price approaches your identified liquidation wall, trimming is far cheaper than betting "the wall holds";
• No big positions in thin-liquidity hours: weekends, late nights, before big events — the same capital moves price much farther;
• Never pick the "prettiest" spot — the prettiest level is also where most orders rest.

---

5. The Right Way to Trade Perpetuals

5.1 Trend Following First; Counter-Trend Bottom/Top Picking Is the No. 1 Killer

The logic of bag-holding spot is "as long as the coin doesn't go to zero, there is still hope" (though vaporware goes to zero anyway); the logic of bag-holding contracts is "three knives: liquidation + funding + ADL" — you cannot afford it.

• The correct profit model for perpetuals = small stops with the trend × big profits; the edge comes from the risk-reward ratio, not the win rate;
• "It already fell 50%, it must bounce" is suicidal logic at 20x: a 50% drop would have liquidated a 20x position four times over;
• Counter-trend trades are not forbidden, but they demand: half size, tighter stops, and objective evidence of "the trend may reverse" (structure break + volume confirmation) — not "I feel like it".

5.2 A Daily Trade-Count Cap

Recommendation: 0~3 trades per day (beginners ≤ 1).

The math: the fixed cost of one round trip = taker fee ~0.05% × 2 + slippage + funding ≈ 0.1%~0.15% per trade. Ten trades a day = 1%+ fixed cost; 20 trading days = 20%+ — high-frequency traders are essentially working for the exchange, and the more trades, the higher the share of emotional decisions.

Frequency	Trait	Verdict
0~1 trades/day	Only planned signals	Healthy
2~3 trades/day	Systematic, disciplined	Acceptable
5+ trades/day	Itchy fingers, revenge trading	Warning sign
10+ trades/day	A fee-burning machine	Stop immediately

5.3 The Weekend and Holiday Liquidity Trap

• Institutions rest on weekends and market depth thins: the same stop order suffers bigger slippage, wicks are more frequent, fake breakouts multiply;
• Crypto runs 24/7 and you do not need to trade every hour — execute planned orders in the most liquid sessions (major trading hours, when institutions are active);
• Weekend practice: cut leverage to 5x or below, or stay flat; open no new positions from Friday night to Monday morning.

5.4 Managing Your State After Consecutive Losses

• 3 losing trades in a row: force leverage down one notch (10x → 5x);
• Down 5% in a day: stop for the day;
• Down 15% in a week: stop for the week; review only, no new positions;
• Account down 30%: treat the principal as halved — treat losses as a real shrink in account size, not as "debt the market owes you".

---

6. Practical Checklists

6.1 The 8 Questions Before Opening

If any single answer fails, stand down:

#	Question	If it fails
1	Direction: is this trade with the trend or against it?	Counter-trend → don't open
2	Timeframe: which timeframe am I entering on, which am I watching?	Can't articulate → don't open
3	Stop: where is the stop? How far, in %?	Not set → don't open
4	Take-profit: is the risk-reward ≥ 2:1?	< 2:1 → don't open
5	Size: is per-trade risk ≤ 2%?	Can't compute → don't open
6	Leverage: is stop distance × leverage < 0.5?	Mismatch → cut leverage
7	Liquidation distance: how far is the liquidation price from market? A wall nearby?	Too close / wall → trim
8	News: any CPI, rate decision, ETF ruling, token unlock in the next 24h?	Big event → trim or don't open

80% of blow-up orders can be filtered out by these 8 questions before entry. After opening, add one self-check: "If this trade hits its stop right now, do I accept it?" If not, both the size and the leverage are wrong.

6.2 The Blow-Up Review Template

Field	Fill in
Date / time
Coin / direction
Entry price / liquidation price
Leverage / margin
Entry rationale
Was it counter-trend at the time
Was a stop set / why didn't it trigger
Was there a liquidation wall near the liquidation price
News backdrop / liquidity session at the time
Direct cause of the blow-up	Counter-trend / oversized / bag-holding / careless stop / wick / funding / slippage
Root cause (one sentence)
The rule to change for next time (one item)

Three review questions:

1. Which of the "8 questions before opening" did this trade fail?
2. Was the loss within-plan or outside-plan? Within-plan losses are tuition; outside-plan losses (removing stops, exceeding size, adding to counter-trend losers) are violations;
3. Which system rule will I change for this trade? — Review is not self-consolation; it is producing one enforceable rule change.

---

7. Self-Diagnosis: Are You a "Contract Gambler"?

7.1 Danger-Sign Checklist (3+ hits = you are in pathological trading territory)

• [ ] Opening a new position to "win it back" within an hour of a blow-up;
• [ ] Doubling down to average losses (Martingale-style "the next one brings it all back");
• [ ] Borrowing money, cashing out assets, or diverting living expenses to fund the account;
• [ ] Bragging about wins, hiding losses, concealing real P&L from family;
• [ ] Manually cancelling a set stop while bag-holding;
• [ ] Using "just one last time" to justify opening a position;
• [ ] Keeping the same leverage — or raising it — after a 50%+ account drawdown;
• [ ] Watching charts 4+ hours a day, with market rhythm hijacking your life.

7.2 How to Exit After Consecutive Blow-Ups

Step	Action	Duration / standard
1. Stop immediately	Close all positions, lock the account (withdraw / change passwords and hand to family), quit all market apps	No logins for at least 72 hours
2. Admit and record	Write every blow-up into the review template in full, no deleting, no sugar-coating	One complete review
3. Cooling-off period	Do only three things: review every blow-up, re-read 01-Perpetual Swaps and this article, practice on the simulator	30 days
4. Spot the "win-it-back" fixation	"Winning it back" is the most expensive trading motive: revenge trading = gambling with leverage. You are clean when you stop thinking about recovery and only ask "does the next trade fit the system"	Continuous self-audit
5. Re-entry threshold	20 consecutive simulator trades executed per plan with acceptable risk-reward before restarting with real money, small size	≤ 2x leverage, ≤ 5% of account
6. Funding red line	Perpetual capital cap = money whose total loss won't affect your life; if you can't meet this bar, never deposit	Permanent

⚠️ Risk Warning

Perpetual trading is one of the few "winner-takes-all, loser-goes-to-zero" games: no circuit breakers, no price limits, 24/7, with leverage, funding, and ADL strangling together. Every checklist and template in this article can only lower the probability of blow-ups, never eliminate the risk itself. If you catch yourself justifying a new position with the words "win it back" or "just one last time", stop immediately, close the leveraged account, or have someone lock the funds for you.

Further Reading

• The Math of Leverage: Why 10x Is Not "10x Profit"
• 06 · Perpetual Blow-Up Case Studies
• 03 · Crypto Derivatives