A trading setup is a written statement of four things: the context in which a trade is worth looking for, the trigger that turns watching into entering, the invalidation price that proves the idea wrong, and the target the idea is aiming at. Each is decided before the trade from a named source: the higher-timeframe bias, a level, a session, an event on the entry chart. Written this way a setup can be taken the same way every time, sized from its stop distance, and tested; a pattern noticed on the day cannot.
A trading setup is a written statement of four decisions made before the trade: the context in which it is worth looking, the trigger that turns looking into entering, the invalidation price that proves it wrong, and the target it is aiming at.
Fourteen parts of this level have supplied pieces — bias and levels, volume, sessions, correlation. This group puts them together into something that can be written on a card, taken the same way every time, and tested. That thing is a setup, and it is not a pattern.
Four parts, in order
A pattern is a shape. A setup is a decision with four parts, and each part answers a question the trade has to answer before it exists.
The order matters for the same reason it mattered in Part 1. Context is decided before a trigger is looked for, so that the trigger cannot invent its own context. Invalidation is decided before size, so that size cannot decide where the stop goes. Target is decided before entry, so that it cannot drift once the trade is open — the target creep Foundations warned about is a target that was never written down.
Context: what must be true before looking
Context is everything that has to be true before the entry chart is opened. It is where most of this level goes, and it is the part most trades skip.
Each line is a part already written. The bias is Step 1; the level is Part 2; the session is Part 8; the dollar’s agreement is Part 12; the count is Part 3. A setup’s context is a list of which of these it requires — not all of them, but the ones it names, and all of those.
Context is also what makes a setup rare. A stack of five filters is true a few times a week on a few instruments. That is not a flaw; a setup that fires every hour has no context, and the overtrading Foundations described is usually a setup with its context removed.
Trigger: one event, not a feeling
The trigger is the single observable event on the entry chart that turns watching into entering. It has to be something two people looking at the same chart would agree happened.
The triggers this level has used are all events of that kind: a close beyond a level, the break of the pullback’s last lower high from Part 1, the third step of the reaction in Part 2, a retest that holds from Part 6. Which one a setup uses, and what each costs, is the next part. What matters here is that it names one.
Invalidation: the price that proves it wrong
Invalidation is the price at which the reason for the trade is no longer true — the swing that would have to break, the level that would have to fail. The stop goes there, with the buffer Foundations described. It is written before entry and it is not moved further away afterward.
Everything about size follows from it, which is the argument position sizing made and Part 14 extended: the risk budget is fixed, the stop distance is set by the invalidation, and the number of lots is what the calculator gives for those two inputs.
The figure’s numbers are illustrative — one per cent of a 10,000-unit account on a pair where a lot is worth ten per pip — but the relationship is not. Which invalidation a setup uses is part of the setup, and it decides how large every trade it produces will be.
Key takeaway Invalidation is chosen for where the idea is wrong, and size is whatever that allows. A setup that chooses its stop to afford a size has reversed the order, and its record will show it.
Target: where it is going, in R
The target is written as a price and as a multiple of the stop distance — the R-multiple from Foundations. It comes from structure: the prior swing, the next higher-timeframe level from Part 2, or a measured move.
Writing the target in R does two things. It tells the setup’s expectancy arithmetic what a winner is worth, which testing will need. And it makes the trade’s quality visible before entry: a setup whose targets are usually under 1R needs a win rate the break-even arithmetic says is hard to sustain, and it is better to know that from the card than from the record.
The card
Put the four parts on one card and it reads like this.
| Part | This setup |
|---|---|
| Context | Daily structure making higher highs and higher lows. Price in a flipped daily level. London open or the overlap. Dollar index not breaking its own support. Three or more reasons. |
| Trigger | 1-hour close above the pullback’s last lower high. |
| Invalidation | Below the 1-hour swing low that preceded the trigger, plus a buffer of one 1-hour ATR. |
| Target | The prior daily swing high. Take only if at least 2R from entry. |
That is a setup. It can be handed to someone else and produce the same trade from the same conditions. It can be sized from its stop distance by the calculator. It can be replayed over a year of charts and given a record, which is Part 18. And it can be refused: the last line of the target row is a filter, and most days it, or the context stack, says no.
A pattern can do none of those things — which is why the record of “trading patterns” is usually a record of a different setup every day.
Next: Entry triggers and confirmation — the three common triggers, and what each one costs in price, evidence and trades not taken.
- A setup has four parts, each decided before the trade: context, trigger, invalidation, target.
- Context is everything that has to be true before the entry chart is opened: bias, level, session, alignment, the count of reasons.
- The trigger is one observable event on the entry chart; a feeling that the level will hold is not one.
- Invalidation is the price at which the reason for the trade is false; the stop goes there, and size is derived from it.
- A setup that cannot be written in a paragraph cannot be repeated or tested.
