forex-scalping.com

How long a losing streak to plan for: the longest run in 100, 250 and 500 trades

Everyone knows losing streaks happen. Fewer people know how long the longest one tends to be, which is the number a risk plan should use.

I simulated 20,000 runs for each case, treating every trade as an independent win or loss. The average longest losing run:

Win rate 60%: 50 trades 3.9, 100 trades 4.6, 250 trades 5.6, 500 trades 6.4
Win rate 50%: 50 trades 5.0, 100 trades 6.0, 250 trades 7.3, 500 trades 8.3
Win rate 40%: 50 trades 6.5, 100 trades 7.9, 250 trades 9.7, 500 trades 11.0

Those are averages. In 1 run in 10 the longest streak was bigger. At 50% over 100 trades the 90th percentile was 8, and at 40% over 100 trades it was 11.

At 1% risk per trade, 11 losses in a row is about -10.5% of the account, since 0.99 to the 11th power is 0.895. A system that wins 40% of the time can do that inside a normal 100 trade stretch, with nothing broken.

Independence is an assumption, and real losses cluster a little more, so treat the table as the optimistic version.

Most talk about losing streaks focuses on revenge trading and overtrading. The opposite problem gets less attention: hesitation. After several stop-outs in a row, the next valid setup arrives and the finger won't click.

The setup meets every criterion. The history says it should be taken. But the memory of the last four losses is louder than the statistics. The order doesn't go in. Price moves in the expected direction, and the winner is missed.

That missed winner does more damage than a loser would have. It confirms the sense that "it only works when it isn't being traded," and the next setup gets even more hesitation.

Hesitation is a form of fear, and fear responds poorly to being told to stop.

A few approaches seem to help. Reducing size for a few trades, so the cost of being wrong feels smaller, while still taking every valid signal. Logging every skipped signal with the result it would have produced, so the cost of hesitating becomes visible. Reviewing the losing streak against the long-term numbers to see whether it falls within normal variation.

The last one carries real weight. A system with a 55 percent win rate will produce runs of five or six losses fairly regularly, and knowing that in advance takes some of the sting away.

Hesitation after a streak isn't a flaw in character. It's a predictable reaction, and it can be managed with structure instead of nerve.

How has hesitation after a losing streak been dealt with in practice?

Useful table. The averages are the starting point, but planning should probably use the 90th or 95th percentile, which are a few trades longer. At a 50% win rate over 500 trades, a run of 11 or 12 losses is not unusual.

What the table implies for sizing is simple arithmetic. If you would be unable to continue after a 10% drawdown, and the 90th percentile streak is 11 losses at 40% over 100 trades, then 1% risk is the ceiling. At 2% risk, 11 consecutive losses would take 0.98 to the 11th power, which leaves 80.1%, a drawdown near 20%. Risking 0.5% brings the same streak to about 5.4%. Pick the size for the streak you can sit through, not the win you hope for. A streak length also depends on how many trades you take, so a scalper with 500 trades a quarter should read the right hand column of the table.

Before applying the table, check your own win rate and trade count over a sensible window. A journal with 140 trades and 44 winners has a win rate of 31%, below the 40% row, so the streaks would be longer than shown: with 31% winners over 100 trades the average longest run is around 10 by the same method. The table is only a guide to scale. You can reproduce it in a spreadsheet in a few minutes with RAND(): one column of 100 rows, win if RAND() is below your win rate, then a formula that counts the longest run. Recalculate fifty times and write down the biggest value.

A correct reading of the 90th percentile is that roughly one in ten 100 trade stretches gives something at least that long. If you run a system for 4 stretches a year, the chance at least one of them reaches the 90th percentile streak is 1 - 0.9 to the 4th, about 34%. So the figure you plan around is not a rare event over a long career. It is closer to a once a year event for a trader with a few hundred trades a quarter. This is why it makes sense to decide in advance what you will do at a streak of 6 or 8, rather than deciding on the day, when the loss column is the only thing on screen.

Streaks of wins also exist, and they are what makes people overconfident. At 50% over 100 trades, the average longest winning run is the same as the losing one, 6, and the table is symmetric. A trader who sees six wins in a row is looking at something that happens in about half of all 100 trade stretches by luck alone. The same is true of six losses. Neither says much about the edge. The statistics only start to speak after a few hundred trades, which is why a streak of any length, good or bad, is a weak reason to change size.