CRT Education · · 6 min read
CRT Setup Invalidation: 6 Signs to Walk Away
When is a Candle Range Theory setup invalid? Six rules I use to cancel a CRT trade, from a Candle 2 close outside the range to a missing turtle soup.
Most Candle Range Theory content teaches you how to find a setup. Very little of it teaches you how to cancel one. That's backwards. Over the years, the trades that hurt me most weren't bad entries on good setups — they were good-looking entries on setups that had already died, and I hadn't noticed.
So here's the direct answer: a CRT setup is invalid when the sequence breaks. Candle 1 sets the range, Candle 2 runs one side into a key level, a lower-timeframe turtle soup proves the run failed, and only then does an entry count. If any link in that chain is missing or happens in the wrong order, there is no trade. Not a weaker trade. No trade.
Below are the six checks I run before I commit risk. If you're new to the model, read what CRT is first and come back.
1. Candle 2 closes outside Candle 1
This is the big one, and it's the rule people bend most often.
The whole idea of CRT is that Candle 2 runs beyond Candle 1's high or low, takes the liquidity sitting there, and then closes back inside the range. That close back inside is what turns a breakout into a purge.
If Candle 2 closes outside Candle 1, the market didn't fail at that level. It accepted price beyond it. That's not a CRT; that's continuation. My CRT Radar treats it the same way: if C2 closes outside C1, the setup is removed from the chart.
What makes this hard is timing. On an H6 or H4 range, Candle 2 can look like a perfect sweep for hours and then close outside in the last few minutes. If you've already entered on a lower timeframe, you're in a trade whose premise just vanished. That's why the turtle soup check below matters so much — it gives you evidence inside Candle 2 instead of a hope about its close.
2. The sweep didn't hit a key level
A random wick through Candle 1's high isn't enough. The purge has to engage a past key level — something the higher timeframe actually cares about.
The key levels I use are order blocks, rejection blocks, fair value gaps, volume imbalances, hidden order blocks, and previous day or week highs and lows. For my Model 1 and Model 2 that means levels from H4 or higher.
If Candle 2 runs the range into open air, with no level there, I have no reason to expect a reversal. The liquidity got taken, sure — but nothing was waiting there to push price back. I walk away.
Notice the wording: engage, not near. A key level sitting a few points above the wick that never got touched doesn't count. On the Radar, the star marks the actual price where Candle 2 engaged the level, not simply the closest unused one.
3. No turtle soup inside Candle 2
This is the rule that saved my account more than any other: no turtle soup, no trade.
A turtle soup, the way I define it, is a candle on the turtle soup timeframe that runs the previous candle's high (or low) and closes back across it. On Model 1 and Model 2 that's an M30 candle. On the Classic 2-6-10 / 1-5-9 model it's M15. I covered why each model uses those pairings in multi-timeframe CRT alignment.
The turtle soup has to print inside Candle 2, at the level it purged. One that prints before Candle 2 starts, or somewhere else on the chart, isn't proof of anything.
Without it, all you have is a wick and an opinion. If you want the deeper background on why failed breakouts work at all, the turtle soup hub walks through it from the beginning.
4. The entry printed before the proof
This one is subtle, and it catches a lot of traders who think they're being disciplined.
Your entry requirements — a purge order block, an inverse FVG, a balanced price range, or a CISD (a close back through the open of the candles that ran into the purge) — only count if they form after the turtle soup.
An order block that formed an hour before the sweep is just an old candle. It might line up beautifully. It might even work. But it wasn't created by the failure you're trading, so it isn't evidence of it.
When I built the CRT Radar, this was the first rule I coded: any entry requirement that prints before the turtle soup is ignored. That single sequencing rule removed most of the low-quality setups I used to take by hand.
5. Price takes out the purge extreme
Once you're in, your invalidation is structural. In CRT, the stop belongs beyond the liquidity sweep — the wick that Candle 2 made when it purged the range. That wick is the whole thesis: "price went there, found nothing, and came back."
If price returns and trades through that extreme, the purge wasn't a failure after all. The market is going where you bet it wouldn't. Get out, or let the stop do its job. Don't widen it.
The same logic applies to an O.T.E. entry. The optimal trade entry zone only arms after a market structure shift and displacement with an FVG in the leg, and the swing origin of that move is the invalidation level. If price closes back through the origin, the structure you were trading no longer exists.
I put risk management first in everything I teach, and this is why. A defined invalidation means you know your loss before you enter. It doesn't mean the loss won't happen. Plenty of my trades end right here, at the stop.
6. The setup has already played out
The last check is the most boring and the most ignored: is the trade still there?
The target in CRT is the opposite side of Candle 1. Once a candle on the CRT timeframe closes through that target, the setup is done. The same goes once Candle 3 closes back through Candle 1. A setup that has already delivered, or already failed, isn't a setup anymore — it's history.
Traders get caught here because a chart full of old boxes looks like opportunity. That's exactly why the Radar clears a setup at that point instead of leaving a graveyard of old drawings from three sessions ago.
A word on news
None of the six rules change around high-impact news like NFP, FOMC or CPI, but the conditions do. The manipulation leg is often bigger and can overshoot the levels you'd normally expect it to respect. Either stay out of those windows or accept that your invalidation sits further away and size down for it. If you trade prop firm accounts, check your firm's news rules too, and see my prop firm guide for how I approach challenges.
My pre-entry checklist
Before every CRT trade, I answer these out loud. One "no" and the trade is cancelled.
- Did Candle 2 close back inside Candle 1 (or is the turtle soup strong enough evidence that it will)?
- Did the purge engage a real past key level from the right timeframe?
- Did a turtle soup print inside Candle 2, at that level?
- Did my entry requirement form after the turtle soup?
- Do I know exactly where the purge extreme is, and is my stop beyond it?
- Is the opposite side of Candle 1 still untouched?
It's not exciting. Walking away from a trade never is. But the setups you cancel are part of your edge, not a gap in it. Rule the candle, rule the trade.
Go deeper
Every one of these rules is written out in full, with the stop and target chapters, in the CRT Handbook. The Handbook is the method; it doesn't include indicator access. If you'd rather the chart enforce the sequence for you, the CRT Radar removes setups when they break these rules, and the free CRT Radar Lite runs the Classic H4 → M15 model with no settings.
Want to see the sequence when everything does line up? Here's the 1:188 MNQ trade at the Sunday Asia open — one example; most of my trades are much smaller or lose. You can see more on the results page.