The analyser: numbers, not advice
This is not an oracle and it is not a bot. It does not predict and it does not recommend.
And the most important thing, because it is what gets misread: it does not promise profits. It will not make a losing trader profitable.
All it has is a broad history of trades, and all it does is put in front of the plan you have drawn how many times price reached the target and how many times it went to the stop first.
It is for ruling plans out. Nothing more, and that is plenty.
How to use it
1. In PLAN → Desk, draw your plan: entry, stop and target. 2. In PLAN → ANALYSIS, press ANALYSE THE PLAN.
It is consulted when you press, not on every tick, because the query leaves the terminal still for a couple of seconds. It reads the history on this machine: nothing leaves the terminal.
If you move the plan afterwards, the analysis stops being valid and the panel says so — THE PLAN MOVED · consult again — instead of showing you a figure for a geometry that is no longer yours.
Which pairs it answers for
It ships a study for eight pairs: AUDUSD, EURGBP, EURUSD, GBPUSD, NZDUSD, USDCAD, USDCHF and USDJPY. No metals, no indices, no other crosses. If you put the panel on another symbol it will say "this pair has no study in your package": that is the correct answer, not a fault.
Each study travels inside the package, with the date it was generated. It is not downloaded or renewed on its own.
The two numbers, and the second one decides
| What it is | |
|---|---|
| Hit rate | Of the times a situation like this one came up, the percentage in which price reached the target before the stop |
| Required | The hit rate you need just to break even, with your target-to-stop ratio and what your broker charges you |
The hit rate on its own says nothing. 35 % is excellent with a target three times the stop, and ruinous with a target equal to the stop. What matters is whether the hit rate clears the bar.
The panel sums it up in a verdict with four values:
| Verdict | What it means |
|---|---|
| THE WHOLE BAND CLEARS THE BAR | Even at the worst end of what was measured, the hit rate clears the bar |
| ONLY THE FIGURE; THE BAND CROSSES IT | The central hit rate clears the bar, but the margin of error crosses it: it could land on either side |
| DOES NOT REACH THE BAR | The hit rate does not get there |
| CANNOT SAY | There is nothing to decide with. It is neither a "no" nor a "yes" |
None of them is "buy", "sell" or a mark out of ten. They speak about the plan against the bar, not about the market.
The required bar may take time to appear
The bar is calculated with your broker's toll: the spread at that time of day and the commission from your own trade history. Without both halves there is no toll.
That is why in a brand-new account you will see the hit rate, but the bar as "no data" until you have at least two closed trades of your own on that symbol: the commission is read from them, and a trade that is open and not closed does not count. It is not a fault: an incomplete toll always lowers the bar, and a low bar passes plans it should not.
And even with the bar in view, the panel warns when it has come up short — "the real bar is HIGHER: slippage not measured" —: slippage cannot be measured, so that number is a minimum. Clear it with room to spare; do not scrape past it.
What you see underneath
| Line | What it tells you |
|---|---|
| Horizon | "within N days, between x % and y %": the hit-rate band that the study declares for that period |
| Expectancy | What is won or lost on average per trade, in multiples of R. It is marked "derived, not measured; no toll": it is worked out from the hit rate and does not include the toll |
| Streaks | The longest losing streak and the winning streak there were |
| Target | Whether your target fits within the median run or goes beyond it (CAREFUL) |
| Stop | Whether your stop holds against what the winners went against you before winning, or falls inside their noise (CAREFUL) |
| Sample | "120 days (340 cases) — the day count is what rules" |
| Pattern | Whether it found the exact situation or had to widen it; and if it gave up a feature, "(dropped: …)", telling you which |
Why the day count is what rules
Several entries on the same day are not several independent observations: it is the same market day counted three times. With too few days, the analyser does not answer.
What the analyser does not know: the news
If there is a high-impact economic release nearby, the panel shows it — time and title — and warns in these words: "This analysis does NOT know about it: it measures what happened in similar situations, and the sample does not separate the ones with a release on top."
And if that kind of release is usually restricted for trading at prop firms, it says so, and adds that yours may not restrict it: check your contract. The panel does not stop you.
When it does not answer — and that IS the answer
| What you see | What it means |
|---|---|
| no data | There was not enough of a sample to say (or the bar cannot be calculated yet) |
| PATTERN NOT RECOGNISED: … | The study does not reach that moment. This is NOT a 0 % hit rate: it is not your plan that fails |
| this pair has no study in your package | It only answers for the pairs that ship with a study |
| not enough candles on this chart yet | Wait until the chart has some history |
| no plan to analyse: draw entry, stop and target | The plan is missing |
| no band to declare: it cannot be said. A gap is not a 0 % | There is no margin that can be calculated |
| WARNING: indicative range | The calibration of that band is not confirmed out of sample: good for orientation, not for deciding |
| no calibration credential: the data does not carry one. Do not take it as good. | The same, said more strongly: nothing vouches for the band |
| The analyser could not read its data on this machine · The analyser gave no result. Press again to retry. | The study could not be read. Press again to retry |
| COULD NOT LOOK: there is nothing to compare this plan against | There is no verdict on your plan, good or bad |
A gap is not a fault. An implicit zero would be: "it never reached the target" and "I had nothing to look it up with" are the best and the worst news possible, and an analyser that said both with the same number would be deceiving you.
What it does NOT do
- It does not recommend. There is no "buy", no "sell", no mark.
- It does not predict. History is not the future, and here it is not dressed up as the future.
- It does not trade. It sends no orders and does not touch your positions.
- It does not score you or learn from you.
- It does not promise profits. It gives you numbers to rule plans out. The decision is yours and so is the result: a good number is not money, and no tool makes a losing trader profitable.
It belongs to the paid version
The free version does not show it. How each version works is in your licence.