DRAWDOWN & THE LOSING STREAK

Losses don’t subtract — they compound. A real SPY drawdown at two position sizes, the recovery table, and why a 10-loss streak is not bad luck: at ordinary win rates, it is an appointment. A drawdown is a hole whose walls are measured in percent, and the ladder out is always longer than the slide in. Slide down 50% and you must climb 100% — double — just to stand where you started, because the climb is measured from the bottom of the hole, where your capital is smallest. Losing streaks are the weather inside the hole: not a sign the sky is broken, just the climate of the activity you chose. You don’t control either. You control one thing only, decided on the surface, before the climb: how much of you is in the hole at all. That is position size — the previous lesson — and this lesson is its evidence.

DRAWDOWN · STAGE 4 · LESSON 3

THE TAKEAWAYS

• The recovery table — pure arithmetic, worth pinning next to the screen: down 10% needs +11.1% to recover; down 20% needs +25.0%; down 25% needs +33.3%; down 30% needs +42.9%; down 50% needs +100%. Nothing in the table is an opinion. Every sizing decision in this course is made in sight of it, because a trader who has to earn +100% to get home does not, in any practical sense, get home.
• The streak table — what ten straight full-risk losses cost, by the setting of Lesson 2’s dial: at 0.5% risk per trade, −4.9% of the account; at 1%, −9.6%; at 2%, −18.3%; at 5%, −40.1%; at 10%, −65.1%. Losses compound geometrically (0.99^10, not 10 × 1%), which is a mercy at small sizes and an executioner at large ones. The same streak, five different fates — chosen in advance by the size dial alone.
• Why the streak is an appointment, not an accident — exact probability arithmetic on the win rate alone: the chance that one specified block of ten trades is all losses is small (about 0.1% at a 50% win rate, 0.6% at 40%). But 200 trades contain 191 overlapping windows of ten, and the chance that at least one of them runs all-loss is about 9% for the coin-flip trader, about 20% at a 45% win rate, and about 38% at 40%. Your intuition prices the single window; your career contains all of them.
• Why this course refuses verdicts before roughly 200 trades — the sample-size arithmetic (standard error of a proportion, first-year statistics): a win rate measured over 20 trades comes with about ±22 percentage points of pure luck either way; over 200 trades, about ±7 points; over 400, about ±5. At 20 trades you cannot distinguish a good method from a lucky one or a bad one from an unlucky one. Judging there isn’t rigour — it is tea-leaf reading with arithmetic on it.

WORKED EXAMPLE — STEP BY STEP

1. Set the episode: SPY’s closing peak of the era, 612.93 on Feb 19, 2025 — then a tariff-driven slide. Thirty-five sessions later, Apr 8, 2025, SPY closed at 496.48. That is −19.0%, close to close. (Intraday, Apr 7 had already printed 481.80 — the hole was deeper inside the days than the closes admit.)
2. Run the fully-invested account: the $10,000 example, 100% in SPY, rode the whole slide: $10,000 × (496.48 ÷ 612.93) = $8,100. It is down $1,900 and needs SPY to gain +23.5% from the trough merely to be whole — the recovery table, wearing a real date.
3. Run the quarter-size account: $2,500 in SPY and $7,500 in cash. The SPY slice fell to $2,025; the account stands at $9,525 — down $475, or −4.75%. Same market, same weeks, same headlines. The entire difference between a −19% event and a −4.75% event was a decision made in February, in calm, about size.
4. Finish the episode honestly: SPY did recover — the first close back above the old peak was 614.91 on Jun 27, 2025, some 56 sessions after the trough. ‘It came back’ is true and useless as a plan: the fully-invested account spent four months underwater needing a +23.5% climb, and nothing in February promised June. In 2022 the same instrument fell −25.4% peak-to-trough (477.71 on Jan 3 to 356.56 on Oct 12) and took far longer about it. Drawdowns end; the ones that end your account are the ones sized wrong.
5. Now the streak walk, same dial, no market required: a 1%-risk trader’s $10,000 after ten straight losses: $9,044 — a bad month, fully described by the plan. A 5%-risk trader’s: $5,987 — a −40% drawdown needing +67% to heal, taken on ten ordinary trades at an ordinary win rate. Neither trader ‘did’ anything unusual. The dial did it.

RISK OF RUIN, IN PLAIN TERMS

Ruin is the drawdown at which you must stop: the account gone, or shrunk below the size where the method can operate, or below the figure your life can carry — Stage 0’s Line 4 number, which is why it was written before any of this. Textbooks compute a precise risk of ruin from a known expectancy, but Lesson 1 already told you the truth: your expectancy is an estimate with wide error bars, so any ruin figure computed from it is a guess wearing a lab coat. What you can do without a model is structural: cap the per-trade risk (Lesson 2), brake the day and the week (Lesson 2’s panel), pre-write the drawdown ladder, and remember that a gap can carry a single loss past its budget (Stage 2) — the framework assumes slippage in the plan rather than perfection in the market. Small size is not timidity. It is what makes the mathematics of survival somebody else’s problem.

COMMON BEGINNER MISTAKES

• Measuring a drawdown in dollars (‘it’s only $900’) instead of in percent of the peak — the recovery table runs on percentages, and it is the percentages that decide whether you get home.
• Assuming a 10-loss streak means the method broke. At a 40% win rate it is roughly a 38%-likely event inside 200 trades: the base rate, not a verdict.
• Doubling size to win it back. The math of the hole punishes exactly this, geometrically.
• Issuing a verdict on a method at trade 20 — a sample whose win rate carries ±22 points of luck. The court isn’t in session yet.

CHECK YOURSELF

DO IT — WRITE YOUR OWN STREAK LINE

In your journal, using your Stage 0 account figure (or the $10,000 example): (1) compute your 10-loss streak cost at 0.5% and at 1% risk, and the account value after each; (2) write your drawdown ladder in full — the action at −10% from your peak and at −20%; (3) write one sentence: what a −19% SPY slide (spring 2025, real) would do to your account at the position size you actually intend to run. These lines transfer verbatim onto the Personal Risk Sheet in Downloads — the gate checks for them there.