A 60% win rate sounds safe until you realize it still hands you a run of seven or eight straight losses if you trade it long enough. Traders don’t blow up because their edge disappeared. They blow up because a perfectly normal losing streak arrived at a size the account couldn’t absorb.
The fix is to size for the streak before it shows up, using math instead of hope.
Why streaks are longer than intuition suggests
Human intuition is terrible at streaks. We feel that a good system “shouldn’t” lose many in a row, so when it does, we assume something is broken and start meddling. But randomness is streaky by nature — flip a coin a few hundred times and you will see runs of six, seven, eight heads with no help from any bias.
Your trading is the same. If each trade is roughly independent, then losing runs are not anomalies to be explained; they are guaranteed features of the distribution. The only real questions are how long a run to expect and whether your sizing can eat it without failing. Treat the streak as a design input, not a surprise.
Estimating your expected max streak from win rate
You can approximate the longest losing streak you should expect over a given number of trades. If q is your probability of losing a single trade (q = 1 − win rate) and N is the number of trades you will take, a workable estimate for the expected longest losing run is:
Longest streak ≈ log(N) / log(1 / q)
Work an example. Win rate 55%, so q = 0.45, over N = 500 trades:
- log(500) ≈ 2.70
- log(1 / 0.45) = log(2.22) ≈ 0.347
- Expected longest streak ≈ 2.70 / 0.347 ≈ 7.8, call it 8
So a 55% system over 500 trades should expect a longest losing run around eight. Note the word expected — the actual worst run can be longer, which is exactly why you plan for a number beyond the estimate rather than right at it.
How streak length scales with total trade count
The uncomfortable property of that formula is the direction it moves. Longer streaks come not from a worse system but from more trades. Because N sits inside a logarithm, the growth is slow but relentless:
| Trades taken (N) | Approx. longest losing streak at 55% win rate |
|---|---|
| 100 | ~6 |
| 500 | ~8 |
| 2,000 | ~9–10 |
| 10,000 | ~11–12 |
The takeaway for anyone trading actively: the more disciplined and consistent you are, the more trades you take, and the longer the worst streak you will eventually meet. A part-time swing trader and a full-time scalper with identical win rates face different worst-case runs simply because of volume. Size for your annual trade count, not for a single session.
Sizing so N losses stay inside your drawdown budget
Once you have a streak number to plan around — take your estimate and pad it, so an estimate of 8 becomes a plan for 10 or 12 — the sizing rule is simple. Your risk per trade must be small enough that that many consecutive full-stop losses still leaves you inside your drawdown limit.
A rough back-of-envelope: if you risk r% per trade and want to survive k losses in a row, your worst-case cluster costs roughly k × r% of the account (a little less with compounding on the way down, but plan with the simple version for margin).
- Want to survive 10 losses inside a modest drawdown budget? Your per-trade risk has to be a small single-digit fraction of a percent to low single digits — not the 3–5% that feels fine on a winning day.
- On a prop-firm account, this is not optional. A trailing maximum drawdown can end the account mid-streak, and the exact rule differs by firm — confirm yours directly. Model the walk-down with a prop-firm drawdown calculator so you can see how a 10-loss run at your chosen risk approaches the breach line.
Once you know the risk percentage, a position size calculator converts it into exact contracts or lots for each trade’s stop distance.
The link between streak survival and risk of ruin
Streak survival and risk of ruin are two views of the same coin. A streak that exhausts your drawdown budget is ruin — the account is gone whether the number that killed it was labeled “streak” or “drawdown.” Both are driven by the same three levers:
- Your edge (win rate and payoff), which sets how often streaks occur and how deep they run. Confirm yours is genuinely positive with an expectancy calculator before sizing anything.
- Your risk per trade, which sets how much each loss in the streak costs.
- Your trade frequency, which sets how many chances the streak has to appear.
Shrink risk per trade and you push both the survivable streak length and the safety margin up together.
Pre-committing to your streak-safe risk level
The whole exercise fails at exactly the wrong moment if it lives only in a spreadsheet. Losing streaks are precisely when discipline erodes — the temptation to “make it back” with a bigger size peaks at loss six, which is also when a bigger size is most lethal.
So pre-commit. Decide your streak-safe risk on a calm day, then remove your own ability to override it mid-streak. Shibiki lets you push that number as a hard limit enforced at the broker, so a run of losses cannot talk you into the oversized trade that turns a survivable streak into a blown account. And because Shibiki’s edge health tracks your live win rate with a Wilson confidence interval, a genuine regime change — as opposed to a normal streak — shows up as your interval shifting, not as a gut feeling at the worst possible time.
Plan for the streak. It is coming. The only choice is whether it finds you sized to survive it.
Related: Trading expectancy · Position size calculator · Prop-firm drawdown calculator