Guides
Guide

How to build an ICT trading checklist

Most ICT checklists fail for the same reason: they're an inventory of everything the trader has ever learned, which makes them unusable in the ninety seconds before an entry. Here's how to build one you'll actually run.

7 min read

Why a checklist at all

You already have a strategy. The problem is that under pressure you don't execute it — you execute a compressed, emotional version of it. A checklist is not there to teach you anything new. It exists to make the decision explicit and repeatable at the one moment your judgment is worst: when price is moving and you want in.

There's a second, less obvious payoff. A checklist turns your trading into structured data. Once every setup carries a record of what was present, you can finally ask which of your confluences are load-bearing and which are decoration. Without that record, you're guessing.

Step 1 — Write down your non-negotiables

Start with the smallest possible list: the conditions that, if absent, mean you simply don't take the trade. For most ICT/SMC traders these cluster around three things:

  • A higher-timeframe bias you've actually defined — and a trade that agrees with it.
  • A liquidity event — a sweep, a raid, a taken pool — rather than entering into untouched liquidity.
  • A shift in structure confirming the reaction, usually a CHoCH/MSS with displacement.

If you can't name three or four of these, that's the real finding. It means your edge isn't specified yet, and no amount of journaling will fix that.

Step 2 — Weight them honestly

Not every confluence is worth the same. An HTF bias that agrees with your trade is doing far more work than "price is near a round number." Give each item a weight that reflects how much it genuinely moves your conviction.

Be suspicious of your own weighting here. Traders reliably over-weight the confluences that are most visually satisfying — a textbook order block — and under-weight the boring structural ones that actually predict follow-through. You'll correct this later with data; for now just record your honest starting guess.

Step 3 — Add your disqualifiers

Disqualifiers are different in kind from confluences. They aren't points in favor — they're gates. High-impact news inside your window. Being outside your prop firm's remaining risk. Trading while tilted after a loss.

A gate that isn't satisfied should stop the trade outright, not merely reduce its score. This is the single most valuable part of the checklist and the part most traders leave out, because it's the part that tells them no.

Step 4 — Cut it until you'd actually run it

Here's the test: could you honestly work through this list with your hand on the mouse and price moving? If the answer is no, it's too long. A checklist you skip is worth nothing, and a shorter list you complete every time beats a comprehensive one you abandon under pressure.

Push the refinements into a second tier. The essentials — the core conditions, the gates, the heaviest-weighted signals — are what you run every time. Everything else is optional detail you consult when you have room to.

Step 5 — Grade before, log after

The checklist only becomes an edge-finding instrument when two things happen in order:

  1. You grade the setup before you enter, committing to a score while the outcome is still unknown.
  2. You attach the result afterwards — R, P&L, or break-even.

That ordering matters enormously. Grading after you know the outcome isn't a checklist, it's a rationalization, and it will quietly teach you the wrong lessons. The whole value sits in the fact that you committed to a number before the market told you whether you were right.

Step 6 — Let the data re-weight the list

After thirty or forty graded setups, patterns start to surface. Some confluence you were certain about turns out to be present in your losers just as often as your winners. Something you almost cut is quietly carrying your best trades.

This is the point of the exercise. The checklist you build today should not be the checklist you run in six months — it should be the one your own results have edited.

Common questions

How many confluences should an ICT checklist have?

Few enough that you complete it under pressure. A practical shape is three to five core conditions, two or three hard disqualifiers, and a second tier of optional refinements you consult when there's time. A short list you run every time beats a thorough one you skip.

Should I grade a setup before or after the trade?

Before, always. Grading after you know the result is rationalization rather than assessment, and it corrupts the data you're trying to build. Commit to a grade while the outcome is unknown, then attach the result once the trade has played out.

What's the difference between a confluence and a disqualifier?

A confluence adds weight in favor of a setup. A disqualifier is a gate — if it isn't satisfied, you don't take the trade regardless of how good everything else looks. News windows, prop-firm risk limits, and trading on tilt are typical disqualifiers.

How many trades before the data means anything?

Treat anything under about thirty graded setups as directional at best. Per-confluence numbers need more still, because each confluence only appears in a subset of your trades. Early patterns are worth noticing and not worth acting on.

Stop guessing which confluences pay.

Grade your setup before you take it, log the result after, and let your own data tell you what's carrying your edge. Free while TradeFern is in beta.

Grade a setup
Keep reading