A signal lands. You read it twice, feel something, and click. The gap between reading and clicking is short, and it is where a great deal of avoidable damage happens.
Most traders try to close that gap by getting better at judging signals. That is the wrong repair. You cannot grade an idea reliably in the ninety seconds after it arrives, because the thing doing the grading is your mood. What you can do is build a gate that every idea walks through in the same order, written down once, before any of them arrive.
This guide is about that gate. Three questions — structure, level, session — written on one card, applied the same way on a quiet Tuesday and a loud Friday.
A filter that can never return a no is not a filter, it is a formality.
Why A Signal Needs A Gate, Not A Reaction
A signal is someone else's reasoning, compressed. It carries assumptions you did not make: a timeframe, a tolerance for heat, a view on where the idea stops being valid, an expectation about how long the position is meant to be held. None of that is wrong. It is simply not yours yet.
Your job is not to grade someone else's idea. It is to decide whether this idea fits a book you can manage. Those are two different tasks, and the second one is the only one you can do well under time pressure.
Without a gate, the same signal gets handled differently depending on the week you have had. After two losers, size creeps up to make the money back. After a good run, size creeps up because you feel sharp. After a stretch of boredom, an idea that would normally fail your own standards suddenly looks acceptable. The signal did not change. The handler did.
That inconsistency is the part of a trading account that is hardest to see in a profit column, because it hides inside decisions that each felt reasonable at the time. A written filter makes it visible. When every trade records which questions it passed, a month of entries starts to show you where your money is leaking — not in the strategy, but in the moments the strategy was ignored.
There is a blunt test for this. If you cannot say what rule you applied, you did not apply one.
GATE, NOT REACTION
A gate
Three questions, written down once, applied the same way on a quiet Tuesday and a loud Friday.The same order every time, so your journal records a rule instead of a mood.
A reaction
You read it twice, feel something, and click.The grading is done by your mood in the ninety seconds after the signal arrives.
The Three Questions Worth Writing Down
The filter does not need to be long. Three questions cover most of the ground, and they are in this order for a reason.
First, structure. Does the direction agree with the structure I already read, on the timeframe I actually trade? This one requires you to have a read before the signal arrives. If your only view of the market is the view inside the signal, you do not have a second opinion, you have an echo. Five minutes with your own chart before the session is what makes this question answerable at all.
Second, level. Is price at a level, or somewhere in the middle of nowhere? Entry location is not decoration. It sets the distance between your entry and the point where the idea is wrong, and that distance is what determines your size. Analysis usually names a zone and a point beyond structure where the premise breaks. If you enter after price has already left the zone, you have kept the full invalidation distance and thrown away the location that justified it.
Third, session. Will this play out in hours I am actually at the desk? An idea that needs managing at 02:00 while you are asleep is a different trade from the same idea in your own session, even though the chart looks identical. Scheduled releases belong to this question too. If a data print or a central bank speaker lands inside the window you intend to hold, the position is exposed to something your reasoning never modelled.
Three questions, three honest answers. Nothing about targets, nothing about how the idea feels.
Writing The Filter So It Fits On One Card
Write it once, when no position is open and nothing is moving. Keep it to four to six lines, because a filter you cannot recall from memory will not survive a fast market.
Phrase every line as a question with a yes or a no. Avoid scoring systems, avoid "mostly", avoid anything that lets you negotiate. A scale from one to ten is just a slower way of doing whatever you wanted to do anyway. Binary questions remove the negotiation.
Add two mechanical lines underneath the three market questions. One for risk: the size you use when every answer is yes, fixed in advance as a percentage of the account, not chosen in the moment. One for management: where the invalidation sits and whether you will be present to act on it. Those two lines turn the filter from an opinion into an order ticket.
Then date the card. Review it monthly against your journal, not against your last trade. If a question never changes an outcome, it is decoration and you can cut it. If you keep losing on the same kind of entry, there is a question missing. Change the card deliberately, between sessions, in writing. The card does not change while a position is open.
HOW THE CARD IS BUILT
The No Trade Branch Is Part Of The Checklist
Most people build a checklist that only knows how to say yes. Every question has a generous reading, and after a minute of interpretation the idea passes. A filter that can never return a no is not a filter, it is a formality.
The awkward case is the partial pass. Structure agrees, the level is clean, but the move is scheduled for a session you will not be at the desk for. Two out of three feels close enough. Decide now, in writing, what you do with two out of three — because deciding it live, with the chart moving, is not a decision, it is a mood. The simple version is that a partial pass is a no, and a question you cannot answer is also a no.
A no is not a loss. It is an outcome your record should keep. Write the signal down, write which line failed, and leave it. Some of those ideas will go on to do exactly what they said they would, and you will have to sit with that. Watching a move you filtered out is the price of having a filter at all, and it is cheaper than the alternative, which is having no gate on the days it matters.
WHAT THE GATE RETURNS
An idea arrives and you run it through structure, level and session.
All three answers are yes.
ThenAct at the size already written on the card, with invalidation placed before the order.
Two out of three, or a question you cannot answer.
ThenTreat it as a no. Write down which line failed and leave it.
Running MMFX Signals Through Your Own Filter
It is analysis, not instruction, and it does not predict anything.
The MMFX daily analysis on XAU/USD is published ahead of the session and is built to be read this way. It maps the structure it is working from, names the zones and the levels that matter, states where the premise would be invalidated, flags the scheduled economic events and the times they land, and carries a conviction reading for the day. It is analysis, not instruction, and it does not predict anything.
Run it through your three lines. On the level question and the event question, the analysis has already done work you would otherwise do from scratch — the zones are marked, the invalidation is stated beyond structure, the release times are listed. Your job there is to check the answer against your own chart rather than accept it.
On the structure question and the session question, nobody can answer for you. Whether the direction fits the read you did this morning, and whether you will be at the desk in the hours the idea needs, are facts about you. That is where the filter earns its place, and that is where most ideas will fail it.
Nothing in the member app can act on your behalf. The AI Trading Assistant is read-only: it cannot place trades, close them, or move money. Orders are placed by you, in your own brokerage account, with money that stays yours.
Trading carries risk of loss. Every decision is yours, and nothing on this page is financial advice.