Losing Streaks in Trading: How Long Is Normal, and What They Actually Cost

You are six losses deep. The strategy worked last month. Nothing about the market looks different. The question in your head is not whether to size down or take a break, it is more basic than that: is this normal, or is this the part where I find out the edge was never there?

That question has an answer, and it does not require a mentor or a gut call. It requires one line of arithmetic that almost nobody runs, because the number it produces is uncomfortable in a way that has nothing to do with discipline.

How Long Is a Normal Losing Run?

Assume for a moment that your trades are independent of each other, which is the assumption worth checking later in this post. Then the chance that any particular stretch of five trades is all losses is simply the loss rate multiplied by itself five times.

The number that matters is not that probability though. It is how many chances the market gets to produce that run. Over 200 trades there are roughly 200 starting positions, and a fresh losing run begins when a win is followed by losses. So:

expected number of losing runs of length k or longer = n x (win rate) x (loss rate to the power of k)

where n is your number of trades. That is the same construction used for the five-loss arithmetic in how to pass a prop firm challenge, applied here without any firm's rules attached.

Run it across four profitable strategies over 200 trades. Each row is a real style, and each one makes money:

The last column is where the expected count falls below one, which is the honest definition of "the longest run you should plan on seeing."

Check the expectancy column before you argue with the table. At 40% and 1:2 you make 0.20R a trade, which is 40R over those 200 trades. That strategy is not merely fine, it is good. And it will hand you a run of eight or nine losses inside a single year of trading, on average, once.

Read the first row next to the last one. The scalper who loses six in a row is at the edge of what their win rate produces. The runner who loses six in a row is not even halfway to their normal worst stretch, and has no information at all.

Why the Worst-Looking Streaks Belong to the Best Strategies

The table has an ordering in it that is easy to miss. As the reward multiple rises, the expected longest run rises with it. The strategy with the highest expectancy in the table, the 1:5 runner at +0.50R, produces the longest runs of red by a wide margin.

This is not a flaw in the approach. It is arithmetic: a target further away is reached less often, so the loss rate is higher, so long runs become ordinary. Win rate is an output of where you put your target, not a measure of skill, and the losing streak is an output of the win rate. Three numbers, one chain.

The consequence is the part worth keeping. The traders most likely to abandon a working strategy mid-stretch are the ones running the best version of it, because their normal looks like everyone else's catastrophe.

Now put recovery next to it. Divide the depth of the run by your expectancy per trade, and you get the number of trades it takes for the edge to make the money back:

The trader with the longest normal losing run recovers it in the fewest trades. The trader with the shortest run takes more than twice as long. The streak that feels worst is the cheapest one on the table, and the one that feels survivable is the one that takes two months of trades to undo.

Both columns assume the edge is real and that trades are independent. That is the whole reason the rest of this post exists.

A Streak Is a Count. A Drawdown Is a Sum.

These get used interchangeably and they are not the same statistic.

Take ten trades of a 1:2 setup, six losses and four wins, in two orders:

• L L L L L L W W W W nets +2R. Longest losing run: 6.
• L L W L L W L L W W nets +2R. Longest losing run: 2.

Identical trades, identical money, and one of them would have had you questioning the strategy while the other would have felt like an ordinary week. Run length carried no information about the result.

The reward multiple pulls them apart further. Five losses is 5R deep whatever your style, but erasing 5R takes five winners at 1:1 and one winner at 1:5. So the same run length means a different recovery burden per trader, which makes comparing your streak against someone else's meaningless unless you also know their target. If you are not already measuring trades in R-multiples rather than dollars, the comparison is worse than meaningless.

The practical version: the run is the number you feel, the drawdown is the number that closes the account, and only one of them belongs in a risk rule.

What the Run Costs, and What Costs More

The run itself is bounded and boring. Eight losses at 1% risk is 8% of the account. You can look that number up before it happens, and if it is unacceptable then your position size was wrong, not your strategy.

What is not bounded is what the run does to the next forty trades. Three things change after a stretch of losses, and each leaves a different fingerprint in the journal.

Size drifts up. The classic. It shows as a rising average dollar loss while the average R loss sits at exactly -1.0R, because dividing by your own risk hides the fact that your own risk grew. An R-only column is structurally blind to this, which is the whole argument of the R-multiple post.

Size drifts down. The quieter one, and it costs about as much. You take the 6R run at full size, then trade the recovery at half size because you have lost confidence. The 6R hole was dug at 1% risk, so it is 6% of the account. The edge earns 6R back, but at 0.5% risk, so it returns 3%. The R column reads flat and the account is down 3%. Nothing in a per-trade R record will ever show you this.

Standards drop. Fewer conditions met per entry, more trades per day, a hunt for the trade that fixes the number. This is the frequency axis and it has its own test, worked in overtrading. The short version is that the trade count is the symptom and the state at entry is the cause.

All three are measurable. None of them are measurable by counting how many losses you have had in a row.

The Stop-After-Three Rule, Priced

Almost every trader eventually adopts some version of "three losses in a row and I am done." It feels like risk management. Here is what it costs on the 40% at 1:2 strategy from the table.

A stop after four consecutive losses fires whenever a run of four or more begins, which is 80 x 0.1296, about ten times per 200 trades. Suppose each firing sits you out for the next five trades. That is roughly 52 of your 200 trades skipped, a quarter of your year.

If those skipped trades are ordinary trades, they were worth +0.20R each. Ten firings times five trades times 0.20R is 10.4R forgone, against a strategy that makes 40R over those 200 trades. The rule deletes about a quarter of the edge, and it prevents nothing, because the four losses that trigger it have already happened.

Now refuse that conclusion, because it rests on the independence assumption stated at the top.

If a losing run reliably changes how you trade, the trades after it are not ordinary trades and they are not worth +0.20R. They are worth less, possibly much less, and the stop is buying something real. The rule pays for itself exactly when post-streak trades average below zero.

So the question is not whether to have a circuit breaker. It is whether yours is earning its 10R. And that is answerable from your own journal:

Tag every trade with the number of consecutive losses immediately before it. Average the R result at 0, 1, 2, 3 and 4+. If the 3+ bucket matches your baseline, your stop rule is a tax. If it is materially negative, the stop is the cheapest fix you have.

Two things make that number honest. Count enough trades in the 3+ bucket to say anything, which by the table above is not many trades per year, so this is a test you run across a full history rather than a quarter. And check average risk per trade in each bucket at the same time, because if size grew during the run then what you are measuring is sizing, not judgment, and the fix is a size rule rather than a stop.

If the answer comes back negative, you have found a state-dependent failure rather than a strategy failure, and the four reasons traders break their own rules covers which of the fixes actually holds while you are in that state.

"Should I Drop the Strategy?"

This is the decision the streak is actually driving, and the streak is the worst possible evidence for it.

Establishing that a 1:2 strategy at 40% has any edge at all takes about 216 trades, derived from the standard-error arithmetic in how many trades before you go live. An eight-loss run is eight trades. It is under 4% of the evidence required, arriving at the exact moment you are least equipped to weigh it.

Three checks beat counting losses, and all three work on the history you already have:

1. Is the run inside the table? Compute your own expected longest run from your win rate and trade count. If your six-loss run sits under a number your strategy produces routinely, stop looking for a cause.
2. Split the sample by date. Compare expectancy before and after the run started. A genuine regime change shows up as a shift across many trades, not as a cluster of consecutive ones.
3. Check the losses individually. Were they your setup, at your size, with your stop? A run of eight clean losses and a run of eight sloppy ones are different events wearing the same number.

Only the third one can be answered at all if your journal holds the fills and nothing else, which is the recurring theme: the number that explains the streak was never in the trade record, it was in the fields you chose to fill in.

One Measurement Trap

Streak counters do not agree on what a break-even trade does, and the choice changes the number on your screen.

In TradingSFX this is the break-even setting. On Ignore, a scratched trade is skipped when the run is counted, so scratching out of a trade does not reset a losing streak. On Win, it does. Six losses with a scratch in the middle is either a run of six or two runs of three depending on a toggle you set months ago. Pick one and know which one you picked before you compare this month against last.

The second trap is scope. A streak counted across your whole history is a different number from one counted inside a single trading account, and if you run more than one, the per-account number is the one that means something.

Running This on Your Own Data

The expected-run table is four multiplications, so a spreadsheet does it. Export your trades to CSV, which is available on every plan including the free one, take your win rate and trade count, and build the row for your own strategy. That is the reference point, and it costs ten minutes once.

For the streaks you actually had, TradingSFX shows two things. The dashboard carries a Current Streak tile, which is the run you are in right now and resets when it breaks, available on every plan. The Win / Loss Streaks chart on Advanced Analytics plots every streak in the filtered set as its own bar, with your maximum winning run, maximum losing run and current run stated above it, so you can read your real distribution against the expected one. Advanced Analytics is on the Premium plan, and there is a live demo with sample data if you want to see the chart before deciding.

The post-streak split is the one piece no journal computes for you, ours included. It is a CSV export, one added column counting consecutive losses before each trade, and a pivot. Worth saying plainly rather than implying a button exists.

Bottom Line

A losing streak is two questions pretending to be one.

The first, "is this normal", is arithmetic, and the answer is almost always yes. Compute your expected longest run once and keep it written down, because the version of you that needs it will not be in a position to derive it.

The second, "what is this doing to me", is the one worth the effort, and it is not answered by the streak at all. It is answered by comparing the trades you took after a run against the trades you took at baseline. If they match, hold your size and keep going. If they do not, you have found something specific and fixable, which is a considerably better outcome than concluding your strategy is broken on a sample of eight.

The streak is not the problem. The forty trades after it are.

Start logging your trades and you will have both numbers by the time the next run arrives.

Published September 25, 2026. Every figure in this post is arithmetic worked openly from the stated illustrative examples, using the independence assumption noted in the text. No statistic here comes from a survey, a study, or measured user data, and no competitor pricing or feature claims are made, so nothing required vendor verification. TradingSFX plan facts and the streak and break-even behaviour described were verified against the application code on 25 September 2026.