A losing streak is rarely unlucky. It is almost always inside the math of your win rate and the number of trades you will take this year. Traders who blow up during one do not fail because the streak was rare. They fail because they sized as if streaks did not exist.
The probability of n consecutive losses
A losing streak is a run of independent (or near-independent) Bernoulli outcomes. If your loss rate is p, the probability of any specific sequence of n losses in a row is:
P(n consecutive losses) = p^nPer-attempt probability of n losses in a row, at three common win rates:
Win rate 3 in row 5 in row 8 in row 10 in row
40% 21.6% 7.8% 1.7% 0.6%
50% 12.5% 3.1% 0.4% 0.1%
60% 6.4% 1.0% 0.07% 0.01%A 50%-win-rate trader hits three losses in a row about one attempt in eight. A 40%-win-rate trader hits five in a row about one attempt in thirteen. Those are not edge-case probabilities. They are normal Tuesday outcomes.
Per-attempt probabilities still do not tell you what to expect over hundreds of trades, though.
How long a streak you should expect over a sample
Take all the trades you will place this year. The longer the sample, the more chances a streak has to surface. The expected longest losing run in n trials with loss rate p is approximately:
Longest losing streak ≈ log(n × (1 − p)) / log(1 / p)Plugging in:
- 100 trades, 50% WR: about 5.6, so expect a streak of 5–6 losses at some point
- 500 trades, 50% WR: about 8.0, so expect a streak of 7–8
- 1000 trades, 50% WR: about 9.0, so expect a streak of 8–9
- 500 trades, 40% WR: about 10.4, so expect a streak of 10–11
- 500 trades, 60% WR: about 6.2, so expect a streak of 6–7
A 50%-win-rate trader who takes 500 trades over a year will, with high probability, see a stretch of 7 or 8 losses in a row at some point. That stretch is the model working as designed. It is not a sign your edge died.
Why streaks feel personal but aren't
Random sequences do not look uniform. Flip a fair coin 30 times and you will usually see at least one run of 4 or 5 same-side outcomes. The brain reads that cluster as something gone wrong, because we expect random to alternate. It does not.
When you are inside a streak, three things happen at once:
- The sequence feels longer than it is, because each loss compounds emotionally
- The loss size feels larger, because recent losses dominate working memory
- The probability of reversion feels lower, because pattern-seeking suggests the streak is the new normal
None of those feelings carry information. They are noise the nervous system layers on top of statistically expected outcomes.
Sizing for the streak you'll hit
The streak you should plan for is not the one you would be unlucky to see. It is the one you are almost certain to see at some point during your trading year. Size so that streak does not end your account.
Capital drawdown after n consecutive losses at risk r per trade:
Drawdown = 1 − (1 − r)^nTwo percent per trade is a common default. Here is what that produces compared to one and three percent:
Risk per trade 5 losses 8 losses 10 losses
1% 4.9% 7.7% 9.6%
2% 9.6% 14.9% 18.3%
3% 14.1% 21.6% 26.3%Now overlay your prop firm's trailing drawdown. Most futures evaluation accounts sit at roughly 4% to 7%. A 2%-per-trade trader who hits a 5-loss streak (statistically expected at any 40–50% win rate, over a few hundred trades) is already at 9.6% account drawdown. The firm's risk system has closed the account before the streak ended. See how trailing drawdown actually works for the mechanism.
Sizing decisions made by trade-level risk alone miss this. Size for the streak you will see, not the trade you are placing.
When a streak is signal, not noise
Streaks fall inside the expected distribution most of the time. They become evidence of edge loss only when they are statistically out of range and you have ruled out process drift.
Two questions before you conclude the strategy stopped working:
- Is the streak inside your model's distribution? Compare to the expected longest streak at your win rate and sample size. A 6-loss streak from a 50%-WR trader over 500 trades is not unusual. An 11-loss streak from a 60%-WR trader, by contrast, has per-attempt probability of roughly 1 in 24,000. That is evidence something changed.
- Did your execution drift? Most strategies do not break statistically. Adherence to them breaks. Audit your last 20 trades against your written rules before blaming the edge. A trading journal that tags rule adherence lets you answer that objectively instead of guessing.
The sample size required to conclude the edge is gone is much larger than most traders think. Six losers in a row at a 50% win rate is a Tuesday. Twenty losers in a row at the same win rate is a 1-in-a-million event per attempt, and that is the territory where you can defensibly say the model is broken. Anything between is your strategy doing what strategies do.
What to actually do during a streak
If the math says the streak is in distribution and your execution audit passes, the response is structural, not emotional:
- Do not increase position size to win it back. That is the single fastest way to convert a normal streak into a blown account.
- Do not decrease position size unless your rules pre-defined a streak-based cooldown. Cutting size mid-streak just locks in a slower recovery once the streak ends.
- Do not change setups mid-streak. You are picking up small-sample bias and abandoning the very edge that produced your expectancy.
- Do log every trade and reread the last 20 entries. If your rule-adherence rate dropped below your normal threshold, the streak is your fault, not the market's.
The traders who survive streaks are not the ones who avoid them. They are the ones who priced them in before the streak started.
Bottom line
A losing streak tests your account size and your rule discipline, not your edge. Compute the streak length your win rate is statistically going to produce. Size below the drawdown that streak would cause. Then execute the same plan you executed before the streak began. Anything else is selling at the bottom of a normal cluster.