Consecutive losses
Also called losing streak, max consecutive losses, consecutive losing trades, losing run
Consecutive losses are losing trades that come one after another with no winning trade in between. The longest such run in a backtest or live record is reported as max consecutive losses.
A losing streak is not proof that a strategy is broken. Even a sound method produces long runs of losses by chance alone. With a 50% loss rate, a run of about 10 losses is normal somewhere in 1,000 trades, and your risk per trade must let you survive it.
What a run of losses tells you that an average hides
Averages describe a strategy over hundreds of trades. You live through those trades one at a time, in order.
Two strategies can lose the same share of their trades and feel very different. One spreads its losers evenly. The other bunches them into runs that drain the account and your confidence.
A common misread is to think a winner is "due" after several losers. The streak itself does not make a win more likely. If anything, a long run can mean conditions have changed, which is why the pause rule exists.
Working through a ten-loss streak in numbers
The figures in this section are an illustrative example, not results from any real strategy. Take a $10,000 account and a strategy where half the trades lose.
Step 1: the chance of a streak starting now. Multiply the loss rate by itself once per trade. Five losses in a row: 0.5 × 0.5 × 0.5 × 0.5 × 0.5 = 0.031, or about 3%. Ten in a row: about 0.1%, roughly 1 in 1,000.
Step 2: the longest streak to expect over many trades. A rare event gets many chances over a long record. You can copy this estimate:
Longest expected streak ≈ log(number of trades) ÷ log(1 ÷ loss rate)
For 1,000 trades: log(1,000) ÷ log(1 ÷ 0.5) = 3 ÷ 0.301 ≈ 10. So a ten-loss run is normal for this strategy at some point in 1,000 trades. Here "log" means the log button on a phone calculator. If you would rather skip the arithmetic, read the expected longest streak from this table:
| Share of trades that lose | 100 trades | 500 trades | 1,000 trades |
|---|---|---|---|
| 30% | about 4 | about 5 | about 6 |
| 40% | about 5 | about 7 | about 8 |
| 50% | about 7 | about 9 | about 10 |
| 60% | about 9 | about 12 | about 14 |
The chart below shows how fast that number grows as the loss rate rises.
Step 3: what the streak costs. Say you risk a fixed 2% of the starting balance on every trade. The sums below show the damage and the gain needed to get back.
If you risk 2% of the current balance instead, each loss is a little smaller than the last. The account ends near $8,171 rather than $8,000.
Step 4: size the risk from the streak. Divide the drawdown you can accept by about 1.5 × the expected streak. The extra half is a safety margin, because real streaks can run longer than the formula. If a 10% drop is your limit: 1.5 × 10 = 15, and 10% ÷ 15 ≈ 0.67% per trade, or about $67. The same ten losses then cost about $670, and a run of fifteen costs about your full 10%.
Keep this margin even when a backtest looks smooth: losses cluster when market conditions change, so trades are not fully independent. Count open trades on linked pairs as one bet, too. Three losers on the same idea in the same hour act like one large loss.
Longest losing run by count vs by money lost
A losing streak can be measured two ways: by how many losses came in a row, and by how much money that run cost.
Those two can point to different streaks. The longest run by count may be many small losses. The costliest run may be shorter but made of bigger ones. Check both.
Is this losing streak bad luck or a broken strategy?
How to find your own longest losing run. Sort your journal by close time and find the longest unbroken run of losers. Compare that count with the expected streak from Step 2.
Mid-streak, the real question is whether to keep going. Use the expected streak from Step 2 as your yardstick. If your backtest shows a much shorter longest run than that estimate, trust the estimate: the test sample is probably too small. Then check two things: how long the run is, and how each trade was taken.
- If the run is within the expected length and every trade followed your plan, keep trading at your normal risk. This is the bad luck you already priced in.
- If the run passes about 1.5 × the expected length, pause. In the example above, that means a 15th loss in a row. If your robot has a pause-after-N-losses setting, use the same number.
- If losses are bigger than planned, pause sooner. Slippage, wider stops or late exits point to execution, not luck.
While paused, compare recent trades with the backtest: same entry rules, same market conditions, same costs. Restart only when you can name what changed and have fixed it, or have confirmed that nothing did.
FAQ
Should I cut my [lot size](/glossary/lot-size) after several losses in a row?
Only if your plan said so in advance. Risking a fixed percentage of the current balance already shrinks each trade as the account falls. Cutting size on impulse usually means the eventual winners are your smallest trades.
Will my live trading match the max consecutive losses in my backtest?
Read max consecutive losses as a stress figure: the worst run the test has shown so far. Treat the backtest figure as a floor, not a ceiling. A longer record gives chance more room, and live costs such as spread and slippage can turn small backtest winners into losers.