Losing streaks

A trader with a 40% win rate will hit a run of eight losses eventually, not because anything is wrong, but because that is what a 40% win rate does.

The question behind the panic

Mid-streak, the question feels like "is my method broken?". It is almost never answerable in the moment, and the attempt to answer it is what does the damage — changing a method mid-drawdown leaves you with no track record for either version.

The answerable version is narrower: is this streak longer than my own win rate would produce by chance?

The arithmetic

For a win rate w, the chance of a run of n losses beginning at any given trade is (1 − w)^n. Across a long sequence, the expected longest run is roughly:

longest expected run  ≈  log(number of trades) ÷ −log(1 − w)

Which produces numbers most traders find surprisingly large:

Win rate Trades Longest losing run to expect
60% 200 about 6
50% 200 about 8
40% 200 about 10
33% 200 about 13

A trend follower with a 33% win rate should expect a run of about thirteen losses in every two hundred trades. Not as a disaster — as the ordinary functioning of their method.

Why this matters more than it looks

Two consequences follow, and they point in opposite directions.

Most streaks are not evidence of anything. If your longest run sits inside what chance produces at your win rate, you have learned nothing about your method by living through it, however bad it felt.

Your position size has to survive the expected run, not the average trade. A method risking 2% per trade with an expected run of ten losses will be down roughly 18% at the bottom of an ordinary streak. That is not a tail risk — it is the base case, and it deserves to be chosen deliberately rather than discovered. The Monte Carlo simulator shows the distribution of outcomes for your own numbers instead of the single average path.

When a streak is actually evidence

Three things distinguish a real problem from a normal run:

It is far outside the expected range. Twenty losses at a 50% win rate is not bad luck; it is roughly a one-in-a-million sequence, and the method is the likelier explanation.

Something changed that you can name. New instrument, new session, new size, or a market regime that shifted. A streak that begins where a change begins is worth taking seriously.

Your execution degraded during it. This is the common one, and it is how an ordinary streak becomes a real problem: the losses were normal, the response to them was not. That shows up as compressed gaps, larger sizes and abandoned stops — see tilt.

What to do inside one

Reduce size, do not stop. Stopping ends the data collection at the worst possible moment. Half size keeps the record continuous while making the outcome survivable.

Change nothing else until the sample is large enough to say something. How large is not a matter of taste — it follows from your own average and variability, and the luck or skill analyzer computes the specific number for your history.

Expect the run before it happens. A trader who knows that ten in a row is ordinary at their win rate behaves differently in the middle of one than a trader discovering it live.

More in Trading psychology, with numbers attached

  • OvertradingWhy volume rises without a decision being made, what it looks like in an export, and the threshold that is worth using instead of an invented number.
  • Revenge tradingWhat happens in the minutes after a loss, why the next trade is different, and how to see it in your own history.
  • TiltWhat tilt is, how it differs from ordinary frustration, and the three things in a statement that give it away.