The same 98 trades went from profit factor 1.26 to 1.78 by moving one number. Why a tight stop gets hunted out of trades that would have worked.
This is the single biggest thing we found, and it is not about entries at all. We changed one number — where the stop goes — and profit factor went from 1.11 to 1.62 on essentially the same trades. Same entry rule, same session, same targets. Better stop.
Our stop sits beyond the swept level, plus a buffer. The buffer was 150 points. We tested it across a wide range, on the same trade list:
| Stop buffer | Avg stop | Profit factor | Holdout PF |
|---|---|---|---|
| 150 points | 385 pts | 1.11 | 1.29 |
| 250 points | 459 pts | 1.14 | 1.45 |
| 350 points | 521 pts | 1.37 | 1.81 |
| 400 points (live) | 546 pts | 1.49 | 1.69 |
| 500 points | 584 pts | 1.62 | 1.81 |
It is not one lucky setting. The whole region from about 350 to 600 points works, across every reward-to-risk ratio we tried, and it holds up on the holdout — data the choice was never fitted to. A single good cell is noise; a broad plateau is a finding.
Think about what you are doing when you fade a sweep. Price has just violently taken out a level. You are betting it comes back. But a level that got swept once is a level that just proved it attracts price — and it very often gets tested again before the reversal actually develops.
A tight stop sits right inside that re-test zone. You get taken out of a trade that then goes on to work, and you pay a full loss for the privilege. The trade thesis was correct. The stop was in the wrong place.
Read that box again, because it is easy to hear "wider stops are better" and simply move your stop while keeping your position size. That is not what this says. That is doubling your risk, and it is how the finding turns into a blown account.
Widening the stop only works if the size comes down with it. If you trade a fixed lot size, this entire module does not apply to you until you fix your sizing first — which is module 6.