02 · Risk Management
An old saying in trading: "Live long enough and the money comes; die too fast and no strategy saves you." This article covers exactly one thing: how to make sure you always have a next hand to play. Decide how much you can lose first, and only then consider how much you can make — that order is the line between life and death in trading.
1. Why Consider Risk Before Reward
1.1 The Asymmetry of Loss and Recovery
First, a mathematical fact: the difficulty of recovering a loss accelerates.
| Loss | Gain needed to break even |
|---|---|
| 10% | 11.1% |
| 20% | 25% |
| 30% | 42.9% |
| 40% | 66.7% |
| 50% | 100% |
| 60% | 150% |
| 70% | 233% |
| 80% | 400% |
| 90% | 900% |
The derivation (50% loss as the example): capital 10,000 → down 50%, 5,000 left → to climb from 5,000 back to 10,000 you need (10,000−5,000)/5,000 = 100% — this is not "a matter of style", it is math.
Conclusion: the defensive value of a trade far exceeds its offensive value. Only after deciding the maximum loss does the potential gain mean anything.
1.2 Three Principles of Risk Management
| Principle | Meaning | Concrete action |
|---|---|---|
| Risk before reward | Before entering, compute "worst-case loss" first, then "best-case gain" | Write the stop-loss first, then the take-profit, on every trade |
| Risk is a budget, not a feeling | Risk is a fixed fraction of total capital, unmoved by mood | Fixed-fraction method (next section) |
| Risk must be measurable and reviewable | "Be careful" is not risk management; "at most 1% of capital lost per trade" is | Log it in the journal; tally it weekly |
2. Position Sizing Methods
The core of position sizing is not "how much to buy" but "how much to lose". Every method revolves around one variable: the fraction of total capital at risk per trade.
2.1 Fixed-Fraction Method (1%-2% risk per trade) — the only recommendation for beginners
Rule: each trade may lose at most 1% (conservative) or 2% (aggressive) of total capital; when the stop triggers, loss = total capital × risk fraction.
Position calculation formula:
Position (amount invested) = risk amount per trade ÷ (entry price − stop price) × entry price
Example: total capital 10,000, risk fraction 1% (= 100 yuan at risk)
Entry 100, stop 95 (5 yuan risk per share)
Shares = 100 ÷ 5 = 20
Position value = 20 × 100 = 2,000 (i.e. 20% of total capital)Why 1%-2%? Look at the damage of losing streaks:
| Risk per trade | Left after 10 straight losses | Left after 20 straight losses | Notes |
|---|---|---|---|
| 0.5% | 95.1% | 90.5% | Very safe, but profits accumulate slowly |
| 1% | 90.4% | 81.8% | Recommended range |
| 2% | 81.7% | 66.8% | Aggressive ceiling |
| 5% | 59.9% | 35.8% | 20 straight losses borders on liquidation |
| 10% | 34.9% | 12.2% | 20 straight losses ≈ out of the game |
Even with a 30% win rate, 1% risk per trade survives a losing streak of nearly 30 trades. The goal of position sizing is not to avoid losses — it is to still be in the game after losing.
2.2 The Kelly Criterion (understand the concept and its limits)
The Kelly criterion computes the theoretically optimal bet fraction given a known win rate and risk-reward ratio:
f* = (bp − q) / b
f* = optimal bet fraction (of total capital)
b = risk-reward ratio (average win ÷ average loss)
p = win rate
q = loss rate (= 1 − p)Numeric example: win rate 55% (p=0.55), risk-reward 1.5 (b=1.5):
f* = (1.5 × 0.55 − 0.45) / 1.5 = (0.825 − 0.45) / 1.5 = 0.375 / 1.5 = 25%Why does almost no one bet full Kelly in practice?
| Problem | Explanation |
|---|---|
| Parameter estimation error | Win rate and risk-reward are estimates; error turns "optimal" into disaster |
| Huge drawdowns | Full-Kelly drawdowns can reach 30%-50%+; most people mentally break mid-way |
| Ignores black swans | Kelly assumes the return distribution is known; tail risk (flash crashes, wicks) is not in the model |
Practical usage: half Kelly or even quarter Kelly. Multiply the Kelly output by 0.25-0.5 before using it. For the vast majority, just use the fixed-fraction method (1%-2%); the Kelly criterion mostly helps you understand "why heavy positions are forbidden" — even a player with a 55% win rate can be wiped out in an ordinary drawdown betting more than 25%.
⚠ 期望 > 0 才谈得上下单；实盘建议从 1/4 凯利起步——凯利假设你输得起，而人输不起。
2.3 Equal-Risk Sizing
Problem: two instruments have different stop distances (BTC stops at 3%, an altcoin at 8%); investing the same amount in each means completely different risk.
Method: make the risk amount equal per trade, not the invested amount:
Risk amount = total capital × 1% (the same every trade)
Quantity = risk amount ÷ (entry price − stop price)
Example: total capital 10,000
BTC: entry 60,000, stop 58,200 (−3%), 1,800 risk per unit → quantity = 100/1,800 ≈ 0.055 BTC
Altcoin: entry 1.00, stop 0.92 (−8%), 0.08 risk per coin → quantity = 100/0.08 = 1,250 coins- Sizing by risk amount makes the risk exposure of every trade in the account uniform; a single failure cannot break you.
- Cautionary counter-example: opening 10 positions at once, each with 10% of capital and a 1% risk budget, is fine; but if each trade puts ~10% of your total capital at risk (heavy size plus a wide stop), five simultaneous losses cost you half the account.
Position sizing quick reference:
| Method | Suited for | In one sentence |
|---|---|---|
| Fixed fraction 1%-2% | Everyone (especially beginners) | Lose at most 1%-2% of capital per trade |
| Half Kelly | Veterans with a proven statistical edge | Kelly output × 0.5, never full |
| Equal-risk sizing | Multi-instrument portfolios | Equal risk amount per trade, not equal invested amount |
| All-in / heavy bets | No one | Mathematical elimination |
3. Stop-Loss Methods
A stop-loss is not "admitting defeat", it is "buying insurance". A position without a stop hands the counterparty a free call option.
3.1 Four Mainstream Stop Methods
| Method | How it's computed | Pros | Cons | Suits |
|---|---|---|---|---|
| Fixed-amount stop | Stop when the loss hits a fixed amount (e.g. 1% of capital) | Simple, precisely computable risk | Ignores market structure; can be swept out by normal noise | Beginners starting out |
| ATR stop | Stop = entry ∓ N × ATR | Self-adjusts with volatility, harder to sweep | Requires understanding ATR; larger losses in wild volatility | Trend following |
| Structure stop | Stop placed beyond the nearest support / resistance | The level has market logic; being swept means structure broke | Stop distance may be large; needs position adjustment | Swing / trend |
| Time stop | Exit if not profitable after N periods | Frees capital and attention; prevents death by a thousand cuts | May miss "slow-start" trends | Intraday / short-term |
3.2 ATR Stop Example
ATR(14) = average true range of 14 candles (example: current ATR = 500, daily BTC)
Long entry at 60,000:
2×ATR stop = 60,000 − 2×500 = 59,000 (stop distance 1.67%)
4×ATR stop = 60,000 − 4×500 = 58,000 (looser; position must shrink accordingly)
When volatility expands (ATR rises to 800):
2×ATR = 60,000 − 1,600 = 58,400 (stop automatically widens, avoiding a sweep)Stop adjustment discipline (iron rules):
✗ Never move the stop down ("give it one more chance" after the loss grows = slow liquidation)
✓ Moving the stop up is allowed (lock profit after gains, e.g. to breakeven)
✓ Execute immediately on touch — no "let me watch it a moment" pauses3.3 Stops Are Not Omnipotent: Three Exceptions
- Gaps / wicks: extreme moves can jump past your stop price (especially with high-leverage contracts); stop slippage can far exceed expectations.
- Liquidity evaporation: small-cap instruments may have no counterparty at the extreme moment; the stop order may not fill.
- Stop distance too small: a stop inside the normal noise band (e.g. 0.5% on a daily chart) gets swept by random wiggles — "over-stopping": shaken out every time, you eventually stop using stops at all.
4. Take-Profit and Trailing Take-Profit
4.1 Two Take-Profit Philosophies
| Philosophy | Approach | Traits |
|---|---|---|
| Target take-profit | Preset a risk-reward ratio (e.g. 1:2, 1:3), exit at the level | Simple, certain cash-out; may sell the big trend |
| Trailing take-profit | Let profits run; exit after a set retrace (fraction or amount) from the peak | Can catch the big trend; gives back more profit |
A common compromise: scale out. At risk-reward 1:1, exit 50% (locking half the profit and lowering the cost basis); trail the remaining 50%.
4.2 Three Ways to Write a Trailing Take-Profit
① Fixed-percentage retrace: exit everything after a 20% retrace from peak profit
Example: cost 100, rises to 130 (peak profit 30%) → retrace 20%×30% = 6 → exit at 124
② Fixed-price retrace: exit after an X% drop (e.g. 5%) from the highest price
Example: high 130 → falls back to 123.5 (−5%) → exit
③ ATR trail: exit after a 2×ATR drop from the highest price (volatility-adaptive)
Example: ATR=3, high 130 → falls back to 124 → exit4.3 The Symmetry Principle of Take-Profit and Stop-Loss
- The stop decides "at most 1% lost on this trade"; the take-profit decides "whether the trade is worth taking" — the ratio of stop distance to target distance is the risk-reward ratio, and it must be computed before entry: trades with risk-reward < 1:1.5 are usually not worth taking.
- Trailing take-profits follow the same "only move in your favor" principle: a trailing level that has moved up never moves back down, or it degrades into "not taking profits again".
5. Maximum Drawdown Control
5.1 The Math of Drawdown and Recovery
| Maximum drawdown | Gain needed | Equivalent consecutive 1%-risk wins |
|---|---|---|
| 5% | 5.3% | 5 trades |
| 10% | 11.1% | 10 trades |
| 20% | 25% | 20 trades |
| 30% | 42.9% | 30 trades |
| 50% | 100% | 50 trades |
| 80% | 400% | 80 trades |
Worked example (capital 10,000, drawdown 50%):
📖 Click to expand: the math of why a 50% drawdown needs a 100% gain to break even
Stage 1: 10,000 → drawdown 50% → 5,000 (lost 5,000)
Stage 2: 5,000 → needs +100% → 10,000 (made 5,000)
Conclusion: the same absolute amount is "−50%" on the way down but needs "+100%" on the way back.
The harsher version: earning it back from 5,000 at 2% risk per trade takes ~35 straight wins.Two practical drawdown-control numbers:
- Daily loss hits 3%-5% → stop trading for the day.
- Account drawdown hits 10%-15% → cut position size (e.g. risk from 1% down to 0.5%) until the drawdown recovers.
- Account drawdown hits 20% → full stop; review strategy and execution — look for the problem, not the market.
5.2 Common Sources of Drawdown (find yours)
| Source | Typical symptom | Countermeasure |
|---|---|---|
| Oversized positions | "This move feels safe" → risk quietly raised to 5%+ | Hard-code the risk fraction; any breach means stop |
| Revenge after losing streaks | A "get-it-back" trade after 3 straight losses | Losing-streak cap + cooling-off period |
| Strategy decay | Running a trend strategy in a range market | Monthly review; pause the strategy once confirmed |
| Black swans | Wicks in extreme markets, liquidity vanishing | Always keep position reserves; avoid high leverage |
6. The Math of Forced Liquidation: Leverage vs Loss Table Most important in this library
🛑 Remember This One: Leverage Does Not Raise Your Win Rate, It Only Accelerates the Road to Zero
This is the single most important table in the entire knowledge base; screenshot it and keep it.
6.1 Leverage vs Adverse-Move Liquidation Table
⚠ 杠杆放大的是伤害不是收益——10x 杠杆下价格反向 10.0% 即爆仓。
Assuming the margin model, no maintenance-margin buffer, wrong direction and refusing to cut:
| Leverage | Adverse move to liquidation | Math |
|---|---|---|
| 1x (no leverage) | Price falls 100% (goes to zero) | Buy with the full account; ride it to zero |
| 5x | ~20% | 1 ÷ 5 = 20% |
| 10x | ~10% | 1 ÷ 10 = 10% |
| 20x | ~5% | 1 ÷ 20 = 5% |
| 50x | ~2% | 1 ÷ 50 = 2% |
| 100x | ~1% | 1 ÷ 100 = 1% |
One-line recap: with 10x leverage long, a 10% adverse move zeroes the principal; for leverage basics see Core Trading Concepts. BTC has moved 10%+ in a single day many times in history — holding a 10x position through it = handing life and death to a day's random noise.
6.2 The Three-Layer Multiplication of Losses
Real loss = adverse move × leverage × position share
Example: total capital 10,000, 50x leverage, fully invested long, price moves 4% against you:
Loss = 4% × 50 = 200% of principal → liquidation, principal wiped out
Same market, only 20% of capital deployed:
Loss = 4% × 50 × 20% = 40% of principal → badly hurt but aliveNote: live contracts also layer on the maintenance margin rate and the liquidation price (see the Futures chapter), and wicks (extreme prices sweeping through in an instant) can fill far beyond the theoretical liquidation price.
6.3 Four Rules for Using Leverage
- Beginners should keep leverage ≤ 3x, used only for hedging or small probe positions; the vast majority of those chasing "get rich on high leverage" end up liquidated.
- Leverage and stop-losses must coexist: leveraged positions without stops are streaking naked.
- Manage position size and leverage separately: high leverage is fine — provided the position is small enough that "fully wrong still loses affordably".
- Nevertop up marginto bag-hold (a margin call is the exchange inviting you to "keep losing").
💀 Never Top Up Margin to Bag-Hold a Losing Position
Never top up margin to bag-hold (a margin call is the exchange inviting you to "keep losing"). Leverage + bag-holding = volunteering for elimination: live contracts also layer on the maintenance margin rate and the liquidation price, and under wick conditions the actual loss can far exceed the theoretical value — liquidation can happen before the percentages in the table.
7. Risk Management Checklist (tick item by item before every order)
□ 1. Have I written the stop price for this trade, with the stop-loss ≤ 1%-2% of capital?
□ 2. Is the position sized from the "risk amount", not a number off the top of my head?
□ 3. Is the risk-reward ≥ 1:1.5? (risk-reward ratio)
□ 4. Does this trade match the entry rules in my trading plan (not a spur-of-the-moment idea)?
□ 5. Is today's / this week's loss budget still intact (the daily / weekly caps from the drawdown rules above)?
□ 6. Am I free of revenge trading, overconfidence, or emotional turbulence?
□ 7. Is the leverage within the cap I set for myself?
□ 8. Do I understand and accept this trade's worst case (slippage included)?
□ 9. Is the journal template ready to fill in right after exit?
□ 10. If this trade loses, will I stop out exactly per plan, no negotiation?
If any answer is no → don't take the trade. There is always another setup; the principal exists only once.Monthly risk audit:
| Check item | Passing line |
|---|---|
| Largest single-trade loss | ≤ 2% risk budget |
| Maximum drawdown in the month | ≤ 15% |
| Rule violations | 0 (target) |
| Liquidations / blow-throughs | 0 |
| Risk-reward ratio | actual average win ÷ actual average loss ≥ 1.5 |
⚠️ Risk Warning
Leveraged trading can wipe out your entire principal and even leave you in debt from a blow-through (some platforms require making up the negative balance). The liquidation percentages in the table are theoretical; actual forced liquidation also depends on maintenance margins, fees, funding rates, and slippage — under wick conditions actual losses can far exceed theoretical values. Everything in this article is for study and research only and does not constitute investment advice; trade derivatives only with money you can afford to lose.