Losing streak
Losing streak — The longest run of consecutive losing trades in a period — the thing that ends accounts, and the thing every win rate guarantees you will meet.
Every positive-expectancy strategy produces losing streaks, and they are longer than intuition suggests. The streak is not a sign that the edge stopped working — it is what an edge looks like from inside, and confusing the two is how people abandon working systems.
In plain English
A losing streak is simply the longest run of consecutive losses in your record. Its importance is that it is the mechanism by which drawdowns actually happen: a drawdown is a streak plus a position size, and neither part causes damage alone.
The length is predictable. At a 50% win rate, the chance of ten losses in a row on any given sequence is about one in a thousand — which sounds remote and is not, because a thousand trades is a few years of ordinary activity. At 35%, ten in a row is roughly one in fifty. Streaks people describe as unbelievable are ordinary output.
The practical consequence is that streak length has to be survivable by design. If a run of ten losses at your current risk would put the account somewhere you would stop trading, the risk is too high — not because ten in a row is likely this month, but because it is certain eventually.
The formula
P(streak of n losses at a given point) ≈ (1 − win rate)ⁿ Expected longest streak in N trades ≈ log(N) ÷ log(1 ÷ (1 − win rate))
- The second line is an approximation and a good one. It says the expected longest run grows with the LOGARITHM of the sample — so more trades means longer streaks, unavoidably.
- It assumes trades are independent. Real trading is not: correlated positions and a single bad regime cluster losses together, which makes real streaks longer than the formula predicts.
At a 50% win rate over 1,000 trades, the expected longest losing run is about 10. At 40%, about 13. At 35%, about 15. Those are expectations, not worst cases.
Worked example — the demo account
The bundled demo account: 96 closed trades at a 51.04% win rate.
| Longest losing run | 4 trades | |
|---|---|---|
| Longest winning run | 4 trades | |
| Win rate | 51.04% | |
| Expected longest run over 96 trades | ≈ 6 | |
| Average loss | $157.44 |
A longest run of 4 against an expectation of about 6 — an ordinary sample, if anything a mild one.
That is the useful reading: nothing unusual happened. A four-trade losing run at a 51% win rate is entirely unremarkable, and the account went on to finish +$4,293.45 up.
The number to plan around is not this one, though — it is the streak this strategy would produce over a thousand trades rather than ninety-six, which is roughly ten. Four consecutive losses at $157.44 each is about $630; ten would be $1,574, or a little over 6% of the opening balance. That is survivable, and it is the calculation worth doing before it happens rather than during.
Every figure above is from the demo account TapeSheet ships with — 96 closed trades, generated from a fixed seed. Open the same account →
What this does not tell you
The caveat is the part worth reading. Most tools put it in a footer, if they print it at all.
- Nothing about whether the edge is intact. A streak is compatible with a working strategy and with a broken one, and the streak alone cannot distinguish them. That determination needs a much larger sample than the streak itself.
- Nothing about the money, because it counts trades rather than currency. Four losses at 1R and four at 2.5R are the same streak and very different events — which is why TapeSheet shows the streak’s total alongside its length.
- It assumes independence, and trading is not independent. Correlated positions lose together, and a regime that suits your strategy poorly produces clusters that no coin-flip model predicts.
- The longest streak in a small sample understates the future. Expected streak length grows with the log of the trade count, so your worst run so far is a floor, not a ceiling.
Where TapeSheet shows it
The Streaks tile on the Overview shows the longest winning and losing runs together with the money each accumulated, because the length alone is only half of the event.
Questions
How long a losing streak should I expect?
Over a thousand trades, roughly ten at a 50% win rate, thirteen at 40%, fifteen at 35%. The right way to use those numbers is to check that a run of that length at your current risk leaves the account somewhere you would still be willing to trade. If it does not, the risk is the thing to change.
Should I stop trading during a streak?
Only if you have a rule that says so, written before the streak began. Stopping mid-run is emotionally reasonable and statistically arbitrary, and its main effect is missing the recovery — which, in a positive-expectancy system, is where the streak gets paid back. A pre-set drawdown limit is a legitimate risk control; deciding mid-run is not.
Does a long streak mean my strategy has stopped working?
It is evidence, but weak evidence, and much weaker than it feels. Deciding a strategy has broken needs a sample large enough to distinguish a change from ordinary variance, which is usually far more trades than the streak contained. What the streak IS good for is telling you whether the losses look like your normal losses — same average size, same instruments, same setups — or like something new.
Related terms
See your own statement, analyzed
Drop an MT4 or MT5 statement on the page and the whole dashboard renders in about three seconds. No signup, no upload — the file is parsed inside your browser tab and never leaves your device.
Analyze my statementAll glossary terms ·Free trading journal ·MT4 statement analyzer ·MT5 report analyzer