Losing streaks with signals: the arithmetic, and the four things that actually help

The 30 days to 21 September 2026 replayed as 79 resolved signals and 3 stops. That is a very good month, and even it had two of its three stops come from signals posted 67 seconds apart in the New York session on 2 September, a EUR/USD and an AUD/USD sell that both failed. Stops cluster, because correlated pairs share the same dollar move. A good month, viewed from inside a bad afternoon, feels like a bad month.

The arithmetic

Plan on a worse rate than the best month. If roughly one signal in eight hits the stop, which is a normal planning figure and not a promise in either direction, two stops in a row are not rare and three in a row is likely somewhere in a thousand signals. Add clustering, because signals on correlated pairs share the same risk on the same day, and a week with four or five stops is something every member will live through, several times a year.

At 1% risk per trade, five stops is about 5% of the account; the risk per trade guide has the same streaks at 0.5%, 2% and 3%. Set that against the record's month, signal by signal, at the same sizing, and a normal streak is a fraction of a normal month. The streak is not the thing that decides your year. What you do in the week after it is.

Why it feels worse than it is

Wins arrive as small, separate events: TP1 here, TP2 there, over days. Stops arrive together, on the one afternoon the dollar moved. The account's equity curve is fine; the emotional curve is not, because the losses are concentrated in memory and the wins are spread thin. Everyone who trades signals has the same distortion. Knowing that does not remove it, but it tells you not to make decisions on the day it is loudest.

The two things that turn a streak into an ended account

Sizing up to recover. Three stops at 1% is 3%. Doubling the size to "get it back" on the next signal makes the fourth stop 2%, and the fifth 2%, and now a normal cluster has cost 7% and the decision to double has been made twice. The signals did not change. The position sizing guide has one formula, and the formula does not have a "recovering" input.

Skipping until it passes. The opposite mistake. The member stops taking signals after the third stop and starts again after seeing two winners go by. They took all the losses and missed the first wins of the recovery, which are, arithmetically, the ones that pay for the streak. Over a year this pattern converts a profitable record into a losing account without the member ever breaking a single rule of sizing.

The four things that help

1. Size so the streak is a number. If five stops in a week would make you do either of the above, your risk per trade is too high. Halve it. A smaller size you keep through the bad week beats a larger one you abandon. The funded account guide shows the same reasoning under a daily loss limit.

2. Decide the rule before the streak. Which take-profit rule, which risk per trade, which signals you skip (news, weekend, late entries). Written down, on a calm day, and not revised inside a bad week. The copier is one way to make the rule harder to break: it applies the same setting to the fourth signal as to the first.

3. Count, do not feel. Keep a journal with every signal taken, its stop distance and outcome. After a bad week, read the month, not the week. If the month's counts look like the record's counts, you have passed one basic consistency check: you took the signals and the outcomes lined up. It says nothing yet about fill quality, sizing, costs or exits, and nothing about whether the month was profitable or the record is right for you. If they do not match, the difference is in your entries, your exits, your sizing, missed signals or costs, and the journal will show which.

4. Check the record, then stop checking. Open the results page, confirm the stops are there with the wins, and close it. Refreshing the equity curve every hour after a losing day is how the two mistakes above get made.

What a good desk does on those days

The same thing. The stop was placed behind the level that invalidates the trade; if price got there, the trade was wrong and the next one is judged on its own. On a day when the dollar is running and the levels are behind price, we post nothing rather than a bad entry; the article on those days explains why the quiet day after a bad one is often the right call. Nobody at the desk doubles the next signal to recover the last one, and the record is the reason: every stop is on it, and the month is still up.

When to stop instead

Same size on the next signal is the rule for a normal streak. Write down, before any streak, the two things that override it: a loss limit for the week or month at which you stop and review (on a funded account the firm's limit does this for you), and an execution problem (fills far from the posted entry, a copier that is not placing, a broker rejecting stops) that stops you until it is fixed. A review is not a change of size; it is a check of whether the rule, the sizing and the execution still match what you planned. If they do, you continue. If they do not, the problem is not the streak.

The short version

Three stops in a row is a normal week. Size so it is uneventful, keep the rule you wrote before it, count the month instead of feeling the week, and take the next signal at the same size unless one of your written pause rules has triggered.