On the Desk’s published record (133 closed trades since 14 Nov 2025), the longest losing streak so far is 9 losses in a row, from 7 Oct 2026 to 14 Oct 2026. At the record’s 44% win rate and its pace of about 151 closes a year, a run of five or more losses happens about three times a year, and seven or more about once a year. Streaks are normal for a method that wins less than half the time; what decides whether you survive one is the size of each loss.
The odds below are the exact run-length maths the record page uses: at a loss probability of 53% a trade, over a year at the record’s pace, the chance of at least one run of each length and how many such runs to expect. They assume each trade is independent; trades that move together, such as several bitcoin miners in one week, make long runs likelier still.
| Run of losses | Chance in a year | How often |
|---|---|---|
| 3 or more | >99.9% | about 11 times a year |
| 4 or more | 99.9% | about six times a year |
| 5 or more | 97% | about three times a year |
| 6 or more | 82% | about once a year |
| 7 or more | 59% | about once a year |
| 8 or more | 37% | about once every two years |
| Losses in a row | From | To | Trades in it |
|---|---|---|---|
| 5 | 15 Oct 2026 | 15 Oct 2026 | MU, INTC, AVGO, AUD/USD, ECHO |
| 9 | 7 Oct 2026 | 14 Oct 2026 | WULF, CLSK, AVAX, BTDR, IREN, TMUS, TMUS, second lot, LUNR, SAIL |
| 5 | 1 Oct 2026 | 6 Oct 2026 | RIVN, EUR/CHF, VERA, EUR/CAD, RHI |
| 7 | 28 Aug 2026 | 8 Sep 2026 | IREN, EUR/CAD, GBP/CAD, NIO, CLSK, USD/CAD, BEAM |
| 7 | 26 Jun 2026 | 22 Jul 2026 | QUBT, WOLF, FLNC, BEAM, NTLA, ASTS, TSLA |
| 3 | 26 Mar 2026 | 30 Mar 2026 | EUR/USD, CORZ, FLNC |
| 3 | 10 Mar 2026 | 13 Mar 2026 | GBP/AUD, CRSP, BEAM |
| 5 | 30 Jan 2026 | 9 Feb 2026 | EUR/GBP, BTC/USD, BTC/USD, BTC/USD, GBP/AUD |
| 3 | 28 Jan 2026 | 29 Jan 2026 | EUR/AUD, GBP/AUD, EUR/NZD |
Most of the long runs share a theme: several squeeze names or several crypto-linked trades failing in the same week. That is how one bad week becomes a long streak, and why the Desk flags trades that are really one bet.
| Risk per trade | A run of 5 | A run of 9 |
|---|---|---|
| 0.5% | about 2.5% | about 4.5% |
| 1% | about 5.0% | about 9.0% |
| 2% | about 10.0% | about 18.0% |
Pick the risk per trade so that the longest run you expect still leaves you calm and inside any prop-firm limit. The losing streak guide walks through the nine-loss run on this record, and the risk per trade calculator shows what each percentage costs.
Every number on this page is computed from the record files when the page is built, with the same rules as the record page: closed trades only, R measured from the original stop as (exit − entry) ÷ (entry − stop), long-term holds counted at the R they booked, and trades in the order they closed. A 0R scratch counts as neither a win nor a loss. When a trade closes on a Sunday update, this page moves with it.
This is one published record, not a law of the market. Past results do not predict future ones, and a different method, market or period will give different numbers.
Every Sunday before the open: analysed swing setups across stocks, forex, gold and crypto with the exact entry, stop and target, an alert when one triggers, the live dashboard and the paid Substack letters.
Get this week’s setupsLonger than most traders expect. On this record, at a 44% win rate, the longest run so far is 9 losses in a row, and a run of seven or more comes about once a year. A method with a lower win rate will see longer runs.
Not on its own. Check that every loss was taken at its planned stop and size. If it was, a run inside the odds above is the method behaving normally; if losses ran past their stops, that is the problem to fix.
With the run-length method: track the chance of being on 0, 1, 2 and so on losses in a row after each trade, and add up the chance of reaching the run length. The calculator on this site does it for any win rate and number of trades.
Most swing traders sit between 0.5% and 1% a trade. At 1%, a run of 9 full losses costs about 9% of the account; at 2% it costs about 18%.