Reduce Position Size During a Losing Streak

Four trades lose, then a fifth, and you’re left holding a question you can’t answer from inside the streak: is this the ordinary clustering that any method with losses in it will eventually hand you, or has the thing that made the method work quietly stopped working? Both look identical from where you’re standing. What helps is that one response covers you either way.

Two explanations, one answer

Say the streak is variance. Smaller size costs you a bit then. The winners in the next batch, and there will be winners, come in at reduced stakes, so the recovery stretches over more trades than it strictly had to. You can live with that.

Now say the edge has gone. Smaller size is the only thing standing between you and a much deeper hole while you work out what broke. Full size on a strategy that has stopped working is how a bad month turns into a bad year, and you usually don’t know which of the two you’re living through until the damage has already been done.

So cutting is cheap insurance against the outcome you can’t rule out, while sizing up to recover is the mirror image of it and, as it happens, much the more common reaction of the two. It assumes the streak is variance and nearly over, which is precisely the thing nobody can know at the time. Get that wrong and your biggest positions sit on your worst stretch of trading.

The instinct runs the other way, obviously. You want it back, and bigger gets it back faster! Anyone who has traded through a bad run knows that pull, and there is a tell for it in other people too: they go quiet. A quiet desk is usually a desk that is sizing up.

Confidence is capital too

Drawdowns do two kinds of damage. The account one is arithmetic and you can read it off the statement. The other is harder to spot because it happens in your head: ordinary setups start looking dodgy, stops go in too tight or too wide depending on the mood, and the next entry comes out either timid or aggressive when the plan asked for neither of those.

Trading small lowers the stakes on all those wobbly decisions. It also makes the next loss something you can absorb. That matters, because absorbing losses is what keeps you in the market long enough to find out whether the method still works. Get scared out of it and you never get the answer.

A worked example

Seven losses running in a trend method, during a market that has turned choppy. All seven near −1R. Each one defensible on its own terms, which is the awkward part.

At full 1% size that run takes 7% off the account. By the eighth trade the journal has started to read differently: entry tighter than the method calls for, size a touch larger than usual, reason column saying “need one.” Of course that trade loses too.

Halve size after the third loss and those same seven trades cost around 4.5% instead, and the eighth probably never gets taken at all, because the pressure that produced it is half as loud.

A simple drawdown ladder

1. Work out what a normal streak looks like for your method. Win about half your trades and five straight losses will turn up regularly over a long enough sample, which on its own tells you nothing at all about whether the method still works.
2. Past that, halve risk per trade at a fixed drawdown from your equity peak. 10% is a common trigger. Apply it until the account prints a new high or a review clears it.
3. At a second threshold, often around 20%, stop and read the journal properly, asking whether the regime has changed, whether the losses are bunched into one setup, and whether they are coming in bigger than −1R, which would point at execution.
4. Restore size on a rule. A new equity high works. So does a stretch of trades at reduced size, expectancy back where it belongs.

The risk management guide covers the drawdown arithmetic that makes these thresholds matter, and thirty trades is not a track record covers why a short run of results, good or bad, tells you less than it feels like it should.

Steps one and four above are each worth more room than a numbered line gives them. How many losing trades in a row is normal? turns the first into arithmetic you can run on your own win rate, so the threshold for “surprising” is a number decided in advance rather than a feeling arrived at on the fourth loss. And when to increase size after a drawdown takes the restoration rule seriously, because it is the half most drawdown plans leave vague, and the half that decides whether a recovery is earned or is simply the next drawdown starting.