cryptotradingbook

Surviving loss streaks

Foundations → Intermediate · ~11 min read

At a glance

  • Loss streaks are statistics, not verdicts - a 50%-win-rate system has a 1-in-8 chance of 5 losses in a row.
  • The damage from a streak is usually behavioural (revenge, oversizing, abandonment), not the losses themselves.
  • Pre-defined streak rules convert a psychological crisis into a routine.
  • Recovery math punishes panic: -30% needs +43%; chasing it back fast needs more than math allows.
  • Stopping is a skill: knowing when the streak is the market vs. when it's you.

1. The math you must accept in advance

With a 50% win rate, a 5-trade loss streak has a ~3% chance per any 5-trade window - but over 100 trades, the probability of some streak of 5+ exceeds 90%. Streaks are not a malfunction; they are the texture of any real edge. A strategy that has never had a 10-loss streak has not been tested, or the record is cooked.

LossGain needed to break even−10%+11.1%−25%+33.3%−50%+100.0%−75%+300.0%−90%+900.0%gain needed = loss ÷ (100 − loss) — beyond −50% the mountain grows steeply
Figure 22.1 — Losses and gains are not symmetric. The deeper the hole, the more force the climb demands.

2. Why streaks break accounts (not just feelings)

The losses are linear; the responses are exponential. Pattern one: revenge sizing - doubling after trade 3, quadrupling after 5. Pattern two: strategy hopping - abandoning the system right before its big trend. Pattern three: stop abandonment - "my stops keep getting hit" leads to trading without them, which converts a drawdown into a liquidation event. All three are documented in journals as the same signature: rule violations clustering after losses.

3. The pre-defined streak protocol

Decide these before the streak: at 3 consecutive losses → size halves until two wins; at 5 → trading pauses for 48 hours and the journal is re-read, not the charts; at 8 or a -15% monthly drawdown → full stop and a strategy review away from the screen. These are not admissions of weakness - they are circuit breakers, and every professional desk has them.

Remember: The streak you handle is the one you pre-decided. Decide your circuit breakers now, write them into the plan, and let the plan be the adult in the room.

4. The two questions that separate streak from signal

First: did the trades follow the plan? If yes, the system is simply in its statistically normal bad patch - changing anything now buys the worst of both worlds. If no - the losses are execution, and the streak is a discipline report, not a market report. Second: did the market regime change? A trend system in a new range loses by design; that is a review item, not a streak.

5. When this fails

Streak handling fails when the circuit breakers are aspirational ("I'll take a break if it gets bad") instead of numeric in the plan, and when recovery is attempted at full size - the "make it back" trade. Recovery is a process: half size until the equity and the adherence stats both recover. The account survives the streak; the ego rarely does, and that is fine.

Risk: The most dangerous moment in any trading career is the day after a painful loss, when the plan says "small" and the feeling says "everything, now". That single moment decides more accounts than any market event.

Knowledge check

  1. What is the per-window chance of 5 straight losses at 50% win rate? (~3% per window - but >50% chance of such a streak somewhere in 100 trades.)
  2. What clusters after losses in journal data? (Rule violations - revenge sizing, hopping, stop abandonment.)
  3. What are circuit breakers in a trading plan? (Numeric pre-defined rules: size cuts, pauses, review triggers at streak/drawdown thresholds.)
  4. What two questions separate a normal streak from a real problem? (Did trades follow the plan? Did the regime change?)
  5. How should recovery sizing work? (Half size until the system re-proves itself - never "make it back" at full size.)