What we learned
Every closed idea gets a post-mortem, and the conclusions become rules the next review has to follow. This page is that record, published in full — including the mistakes.
A loss is not automatically a mistake. Every review scores the process, not the profit. An idea can be built correctly and still lose, because the outcome of any single trade is largely uncertain — and it can win after a genuine error. We grade them separately: mixing the two teaches you the noise instead of the method.
Active rules (2)
Data release overlooked
Occurrences: 2
Sep 10, 2026
A winning trade can still be a rule breach
CS-2026-0003 closed at +2R with a CPI release sitting inside its horizon. The result was good and the process was not: the outcome was decided by the data print, not by the setup. Counting this as a success would teach the method to ignore the calendar whenever the chart looks convincing.
Rule now applied
Check twice: check the macro calendar against the idea horizon BEFORE calculating levels; never publish an intraday idea with a tier-1 release inside its horizon, even when the setup looks clean.
Stop too tight
Occurrences: 1
Sep 10, 2026
The stop belongs outside the invalidation, not inside it
Two closed trades, CS-2026-0001 and CS-2026-0004, both won and both carried the same defect: the protective stop was placed a few pips inside the level published as the invalidation. That creates a band where the position is already closed for a full loss while the thesis is still valid by its own wording. Neither trade was tested in that band, so both wins are uninformative about whether the risk was correctly placed. The order of work was backwards: levels first, invalidation written afterwards to fit them. From now the invalidation price is chosen first from structure, the stop goes beyond it with room for the wick, size follows from that distance, and R:R is computed last. An idea that only clears 1.5 because the stop was pulled in is not an idea that clears 1.5.
Rule now applied
Derive the stop FROM the invalidation level plus a noise buffer BEFORE computing R:R; the stop must sit beyond the invalidation price, never inside it, and if R:R then fails 1.5 the idea is skipped rather than the stop tightened.
Recent post-mortems
| # | Instrument | R |
By process quality | Type |
What changes from now on |
| CS-2026-0004 |
USD/JPY |
3.60 |
Process was flawed |
Stop too tight |
Same correction, and it now applies as a standing habit rather than a one-off: derive the stop FROM the invalidation level plus a noise buffer, and only then compute R:R and targets. Never write the invalidation after the levels to make them fit. On USD/JPY specifically, 50 pips of risk is inside normal noise; the structural stop was nearer 150.45/150.50. |
| CS-2026-0001 |
EUR/USD |
2.00 |
Process was flawed |
Stop too tight |
Place the stop BEYOND the invalidation level, never inside it. If invalidation is an hourly close below 1.0815, the stop belongs under that structure with room for the wick (roughly 1.0805/1.0808), and position size follows from that wider distance. If the resulting R:R no longer clears 1.5, the correct action is to skip the idea, not to pull the stop back in to make the ratio work. |
| CS-2026-0003 |
US 500 |
2.00 |
Process was flawed |
Data release overlooked |
The profit hides a rule breach: the checklist forbids an intraday idea with a tier-1 release inside its horizon, and this one was published anyway. It won on the coin flip. From now on the release calendar is checked against the horizon before the levels are even calculated. |
| CS-2026-0002 |
XAU/USD |
-1.00 |
Process was sound |
No process error |
Nothing. The invalidation level was correct and the loss was one unit of risk, exactly as planned. Being stopped out before the move is the cost of having a defined stop, not an error to correct. |
Where the method is actually working
This is a small sample. Read every pattern here as a hypothesis under test, not as a proven edge — and be sceptical of any service that claims otherwise.
By asset class
| Closed |
Win rate | Avg R |
|---|
| forex |
2 |
100% |
2.80 |
By direction
| Closed |
Win rate | Avg R |
|---|
| LONG |
3 |
100% |
2.53 |
By review
| Closed |
Win rate | Avg R |
|---|
| Midday review |
4 |
75% |
1.65 |
By horizon
| Closed |
Win rate | Avg R |
|---|
| intraday |
2 |
100% |
2.00 |
| swing |
2 |
50% |
1.30 |
By conviction
| Closed |
Win rate | Avg R |
|---|
| 4 |
2 |
100% |
2.80 |
By exit
| Closed |
Win rate | Avg R |
|---|
| Target 1 hit |
2 |
100% |
2.00 |
By error type
| Closed |
Win rate | Avg R |
|---|
| Stop too tight |
2 |
100% |
2.80 |
By process quality
| Closed |
Win rate | Avg R |
|---|
| Process was flawed |
3 |
100% |
2.53 |
Method & limits →
Track record →
⚠ Risk Warning. Trading CFDs and other leveraged products carries a high level of risk and can result in the loss of all your capital. Only trade money you can afford to lose. General information, not investment advice.