04 · Equity Curve and Performance Attribution

The first three articles of this chapter covered "how to write a plan, how to control risk, how to manage your mind"; this one is the acceptance test: whether your system works is not decided by a handful of winning trades but by one thing — the equity curve. It also answers the sharper question: when the curve rises, is it because the system is good, or because the market is good? Where exactly does the money come from? That is performance attribution.

1. What the Equity Curve Is: A Trader's "Health Report"

1.1 Definition of the Equity Curve

Equity curve: the account equity (principal + floating P&L) plotted as one line over time. Record a point after every trade closes and after every daily close; the points joined together form your account's "electrocardiogram".

Why call it a "health report": a health report doesn't look at a single blood-pressure reading; it looks at trends and abnormal ranges. Same for the equity curve —

1. Single-trade P&L is noise; the curve is the signal. Winning 3 in a row doesn't mean the system is good, losing 5 in a row doesn't mean it is bad; only the shape and slope of a curve built from enough samples carry information.
2. The curve is honest. You can tell yourself "I'm actually profitable, the market has just been bad lately", but the curve doesn't lie — it reflects the net result of money in and money out.
3. The curve exposes execution problems. A stretch of abnormally steep decline usually maps to violations or revenge trading in that period (verify against the journal).
4. The curve is the system's "blood pressure + blood sugar + blood lipids" combined report. Slope shows earning power, drawdown shows risk control, volatility shows style stability.

Core idea: you are not trading a handful of market moves; you are trading "this curve". Before every order, ask: does this trade help the curve's long-term shape, or does it just add one more random jitter?

1.2 Plotting Tools

• Beginners: manual entry in Excel / Google Sheets — one "equity" column plus a line chart is fully sufficient (template in Part 7).
• Advanced: a script pulling historical P&L from the exchange API to update automatically (see the Quant Practice chapter↗).
• Key requirement: record continuously, don't cherry-pick. Keeping only the profitable months and deleting the losing ones distorts the health report.

2. The Three Health Shapes of an Equity Curve

2.1 Shape Overview

2.2 Breakdown

① Steady uptrend — the ideal, and rare. Traits: shallow drawdowns (<10%), quick recoveries, stable slope. Note: such curves usually come with "a mediocre win rate but an excellent risk-reward ratio" — frequent small stop-losses + occasional big wins mean the curve climbs slowly most of the time.

② Spike drawdowns — the most common problem account. Typical script: three months of small gains → one heavy-position month → all profit given back. The causes of the spikes are almost always the same few:

③ Long flat stretch — the biggest confidence drain. Flat ≠ losing, but 3+ months of flat usually means one of two things: the strategy has no edge in the current environment (a trend strategy in a prolonged range market), or your execution is steadily bleeding away EV (violations grinding positive expectancy down to zero). The test is the compliance rate: flat with compliance > 90% → a strategy problem, change the strategy; compliance < 70% → fix the human first, don't rush to change the strategy.

In one sentence: the curve's shape is the system's diagnosis, and both spikes and flats need "triage before treatment" — first decide whether it is a strategy disease or an execution disease, then act. (The triage method is in Part 5, performance attribution.)

3. Key Performance Metrics: Review and Deepening

3.1 Metric Quick Reference

These numbers are common-sense reference lines, not promises. Anyone promising you "50% annualized, 5% drawdown, Sharpe 3" belongs on your blacklist — that is a photoshopped curve.

3.2 How to Read Contradictions Between Metrics

There is no "perfect across the board" metric combination; the combination itself describes a style:

Scenario A: high Sharpe, low return. Say 12% annualized, 6% drawdown, Sharpe 1.8. Reading: very steady but earns too little — most likely a range / grid-type strategy, high win rate, small per-trade results. Suited to the stability-minded, but note: high-Sharpe systems often have small capacity and fail the moment the regime changes (one-way trend).

Scenario B: high return, high drawdown. Say 40% annualized, 30% drawdown, Calmar 1.3. Reading: it earns, but the ride is agonizing — most likely a trend-following strategy, long flat stretches + occasional eruptions. Suited to those who can bear the psychological pressure, but stress-test with the drawdown number: a 30% drawdown means 100k becomes 70k — could you keep executing the plan in that state?

How to choose:

Core idea: metrics are style filters, not report cards. When forced to choose between high Sharpe and high return, pick "the one that lets you sleep" — only the sleeping endure.

4. The Psychology and Math of Drawdown: The Recovery Timetable

4.1 The Math of Drawdown and Recovery

After a drawdown of n%, a gain of x% is needed to break even: x = n / (1 − n). That is why "controlling drawdown" is not a style question but a math question:

4.2 Recovery Timetable (at 3%, 5%, 8% Monthly)

Three brutal corollaries of this table:

1. After a 50% drawdown, even steadily earning 5% a month takes 14 months to break even. For that year-plus you are barely better off than someone starting from zero — drawdown destroys not just money but time.
2. "Keep drawdown within 20%" matters because it is the floor for "recoverable within a year". Beyond 30%, the combined psychological and time cost makes most people systematically give up.
3. Drawdown damage is not linear. 20%→30% looks like just 10 points more, but recovery time nearly doubles. Before every size-up, think how far you stand from "a wasted year".

The psychological point: the scariest part of a drawdown is not the money; it is how it keeps corroding your execution during the "recovery period". That is why the Trading Psychology article stresses: keep positions small enough that "losing doesn't hurt" — mathematically the drawdown stays controllable, and psychologically you can survive the repair.

5. Introduction to Performance Attribution: Where the Money Was Made and Lost

5.1 Why Attribution Matters

The equity curve answers "did you make money"; attribution answers "on what basis". Traders who skip attribution credit themselves in wins and blame the market in losses — never learning what their edge is, and thus never able to amplify it or plug the leaks.

5.2 Dimensions of Decomposition

5.3 Example P&L Attribution (Monthly Attribution Table)

Below is a fictional account's monthly attribution (42 trades, net +8,460 USDT):

Three lessons from this table:

1. The money was made by "strategy A + instrument X + session Y + long direction"; the other combinations roughly break even or lose — the next step is to contract the battle line and trade only the winning combination.
2. The 6 violating trades contributed −1,840: without them the month would have been +10,300. The cost of violations is real and countable — not a moral lecture.
3. Every conclusion points to an "action": pause the dip-buy strategy, cut SOL, reduce Asia-session trades and shorts, investigate the violations one by one — attribution's output must be an actionable list.

5.4 Common Attribution Mistakes

6. Setting Equity Curve Targets for a "Robust System"

Common-sense targets, not promises, let alone guarantees. Define "good enough" first, and you will know what the curve should look like.

6.1 Reference Specs for a "Robust Curve"

6.2 Why 15%-30% / 20%

• Annualized above 30% with drawdown under 20% is a tiny top-tier combination in real markets; ordinary traders who target it are presetting failure.
• Annualized below 15% doesn't pay for the time and psychological cost of active trading — an index fund beats it.
• The 20% drawdown floor comes from Part 4's recovery table: 20% needs +25% back, about 4.6 months at 5%/month — that is a "bearable repair cycle".

The point of a target is not "hitting it"; it is giving the curve a frame of reference. Check monthly: is the drawdown closing in on 20%? Is the annualized slope enough? If the direction is wrong, change course that same month.

7. Equity Curve Management in Practice

7.1 Daily Logging Template (10 minutes)

# Equity curve log (2026-08-16)
- Account equity: ____ USDT (____% vs yesterday)
- Month to date: ____% (same period last month: ____%)
- Maximum drawdown this month: ____% (cap 20%; breaching triggers the circuit breaker)
- Trades closed today / P&L: ____ trades / ____ USDT
- Floating P&L of open positions: ____ USDT (not realized; logged separately)
- Violations today: □ none □ yes (____)
- One line: ____ (e.g. "executed strictly today, nothing unusual")

7.2 Monthly Review Report Template (1 hour)

# Monthly review report (month ____)

## 1. Numbers overview
| Metric | This month | Last month | Target |
|---|---|---|---|
| Return | | | 15%-30% annualized, monthly equivalent |
| Maximum drawdown | | | ≤ 20% |
| Win rate / risk-reward | | | — |
| Number of trades | | | ≥ 30 samples for statistical meaning |
| Violations | | | 0 |

## 2. Equity curve shape
- This month's shape (steady uptrend / spike drawdown / flat): ____
- Chart: paste / describe the key turning points: ____

## 3. Performance attribution (see the table in Part 5)
- Main profit source this month: ____
- Main loss source this month: ____
- Next month's action list: ①____ ②____ ③____

## 4. Drawdown and risk review
- Deepest drawdown this month: which day, and why: ____
- Days to recover: ____

## 5. Execution and psychology review
- Violation list and the emotion at the time: ____
- Emotion-outcome cross-table (see Trading Psychology, 6.3): ____

7.3 When to Stop (execute on trigger, no negotiation)

Stopping is not surrender; it is the highest-level move in equity curve management. Every stop-loss and every stop serves the long-term shape of the curve.