How Many Losses in a Row Is Normal? Work It Out from Your Win Rate

Losing several trades in a row is normal for every system, and you can work it out in advance from the win rate. A system that wins 40% should expect a losing streak of around 8 trades at some point over 100 trades. A system that wins 20% can lose more than 20 in a row without anything being wrong with it. This article does the math step by step, then shows why short-term results are driven by luck more than you might think.

If win rate and RR are still fuzzy, read Win rate and breakeven alongside this.

The chance of losing N trades in a row

Let w be the win rate (the share of trades that win). The chance of losing one trade is 1 − w. If each trade is independent of the others, the chance of losing N trades in a row is that number multiplied by itself N times.

Chance of losing N trades in a row = (1 − w) to the power of N

Example: a system that wins 40%. Chance of losing per trade = 1 − 0.4 = 0.6

• 5 losses in a row = 0.6 × 0.6 × 0.6 × 0.6 × 0.6 = 0.078, or about 7.8%
• 9 losses in a row = 0.6 to the power of 9 = 0.010, or about 1%

That sounds small, but it is only the chance for “this particular set of 5 trades”. Over 100 trades, a losing streak can start at almost any trade. A small chance that gets tried dozens of times is almost certain to happen.

The longest losing streak to expect

There is a rough formula for the longest losing streak when you take n trades in total.

Approximate longest losing streak ≈ log(n) ÷ log(1 ÷ (1 − w))

Use base-10 log, which any phone calculator has. Try it with a system that wins 40% over 100 trades.

1. log(100) = 2
2. 1 ÷ (1 − 0.4) = 1 ÷ 0.6 = 1.667, and log(1.667) = 0.222
3. 2 ÷ 0.222 = 9.0 trades

Check: we already worked out above that the chance of losing 9 in a row = 0.6 to the power of 9 ≈ 1%, or 1 in 100. Over 100 trades there are about 100 starting points to try, so you expect to see it about once, which matches the formula.

The formula is an estimate. We also did the exact calculation by adding up the probability of every possible streak length. The same case gives an average of 7.9 trades, so the formula comes out slightly high, which is the safe side when planning. And the average is not a ceiling: in this case there is about a 20% chance of a losing streak of 10 or more.

Longest losing streak by win rate

Every cell assumes each trade is independent and the win rate stays constant.

Win rate	100 trades (formula)	100 trades (exact)	500 trades (formula)	500 trades (exact)
20%	20.6	15.6	27.9	22.7
30%	12.9	10.7	17.4	15.2
40%	9.0	7.9	12.2	11.0
50%	6.6	6.0	9.0	8.3
60%	5.0	4.6	6.8	6.4

Two things stand out in the table. First, when the win rate drops from 50% to 20%, the longest losing streak gets about 2.6 times longer (6.0 → 15.6). Second, trading 5 times as long does not make the streak 5 times longer, only a few trades longer, because it grows with the log of the number of trades.

A system that wins 20% and lost 21 in a row, yet is still a good system

In our backtests, one system buys pullbacks to the 61.8% level of a large swing on gold (XAUUSD) M1. It was tested from 2024 to September 2026, using 1-minute price data for 2024–2025 and real ticks for 2026. This system wins only about 20%, but its winners average 7.4 times the size of its losers, and it once lost 21 trades in a row.

Let us work out the result per trade in units of R (the money lost when one Stop loss is hit).

• 20 wins out of 100 trades earn 20 × 7.4 = 148R
• 80 losses cost 80 × 1 = 80R
• Net 148 − 80 = 68R over 100 trades, or about +0.7R per trade

This was the highest figure among the systems we ranked, despite having the lowest win rate. Compare it with the table: a system that wins 20% over 500 trades has an average longest losing streak of 22.7, and about a 60% chance of a streak of 21 or more. So losing 21 in a row is something to expect, not a sign the system is broken.

All of these are backtest results. Past results do not predict future results, and these numbers are not returns you should expect.

One good year may just be luck

A losing streak can make a good system look bad. The reverse is also true: a lucky run can make a system with no edge look good.

In our backtest, we opened one gold (XAUUSD) trade a day at 06:30 Thai time (UTC+7), 0.10 lot per trade, starting balance 10,000 dollars, Stop loss 5 dollars (a 50 dollar loss), take profit 30 dollars (a 300 dollar gain), so RR 1:6. 2022–2025 used 1-minute price data and January to September 2026 used real ticks. One year is about 250 trades.

Swings caused by pure luck are measured with the SE (standard error). You can work it out roughly like this, assuming the system wins about 17% (near its breakeven of 14.3%).

1. The result per trade swings between +300 and −50, a 350 dollar gap.
2. Swing per trade ≈ 350 × √(0.17 × 0.83) = 350 × 0.376 ≈ 131 dollars
3. Over 250 trades, the swing grows with √250 ≈ 15.8, giving 131 × 15.8 ≈ 2,080 dollars

So a system with no edge at all can easily have years that are up or down by about 2,000 dollars. And over 5 years (about 1,224 trades) it can still swing by about ±4,500 dollars.

This rule uses the colour of the last closed H4 candle to pick the direction. Look only at 2025 and you see a 4,007 dollar profit, which looks convincing. But the other 4 years lost money, and the 5-year total is just +548 dollars. In this test set we tried 9 direction rules, and the 5-year total of every rule fell within ±2 SE of zero. None of them could be told apart from luck.

Look at it another way. A year that is up by more than 2 SE purely by chance has about a 2.5% chance per year. Try 9 rules × 5 years = 45 cells, and you would roughly expect about 45 × 0.025 ≈ 1 cell with a great year like this, even if no rule had any edge. The more variations you try, the easier it is to find a great year that came from luck. More on this in Peaks vs plateaus when tuning settings and When the market changes regime.

All of these are backtest results. Past results do not predict future results, and these numbers are not returns you should expect.

Choose a risk per trade that survives the losing streak

Once you know how long a losing streak is normal, the next step is choosing a risk per trade that gets you through it. This builds on Why many traders risk only 1% per trade.

Example: a 10,000 dollar account, a system that wins 20%, preparing for a 21-trade losing streak, risking a percentage of the remaining balance on every trade.

• Risk 1%: 0.99 to the power of 21 = 0.810 is left, or 8,097 dollars, a 19% drop
• Check: the first trade loses 100, leaving 9,900. The second loses 99, leaving 9,801. Each trade loses a little less, so the drop is slightly less than 21 × 1% = 21%.

Risk per trade	Account drop after 21 losses in a row	Gain needed to get back to break even
1%	19.0%	23.5%
2%	34.6%	52.8%
3%	47.3%	89.6%
5%	65.9%	193.6%

At 1% it hurts, but you can keep trading. At 5% the account is down to about a third, and you need to make almost three times what is left to break even, even though the system may still be fine. Also ask yourself honestly: if you saw the account drop 19%, could you keep following the system?

The right trade size is one that leaves you with enough money, and enough nerve, to keep trading after the longest losing streak you can expect.

Set your review rules before you start

The middle of a losing streak is the worst time to make decisions, because emotions take over. Write the rules down before you start.

1. Find the system’s win rate from a test with enough trades.
2. Open the table and find the longest losing streak to expect for the number of trades you will take in that period. Use the formula value, which is higher, to be safe.
3. Work out how much the account would drop with that streak at your risk per trade. Can you accept it?
4. Set a review line, for example a losing streak longer than the formula value, or an account drop past a level you chose.
5. When you hit the line, stop opening new trades. Check whether you followed every rule and whether the market has changed from the period you tested. Then decide.

If you have not hit the line, the losing streak is a normal cost of the system, not a reason to change the rules.

Common mistakes

• Dropping a system mid-streak while it is still within the expected range and starting a new one, which will have losing streaks of its own.
• Increasing the lot after losses to win it back. Long losing streaks are exactly what breaks this approach fastest. See Grid, Martingale and averaging down.
• Judging a system on a few dozen trades or a single year. Short-term results contain a lot of luck, both good and bad.
• Thinking “I have lost so many in a row, the next one must win”. If each trade is independent, the chance of winning the next one is still the same.