Myth Test #17
Does Candle Range Theory (CRT) work? We tested 3,383 NQ setups
Candle Range Theory (CRT) is everywhere on trading Instagram: one candle sets a range, the next sweeps one side and closes back inside, and you trade to the other side. We tested it on NQ and gold, from 1-hour to daily candles.
Short answer
Not in our test. On NQ futures 4-hour candles from April 2017 to September 2026, CRT took 3,383 trades with 49% winners and lost −$37K on one contract (−$11 a trade after costs). Taking the same trades in a random direction lost less (−$5 a trade). Every version we tried lost money on newer data: 1-hour, 4-hour and daily candles, on NQ and on gold, with trend filters and with only the 1, 5 and 9 AM candles.
The rules we tested
- Candle 1 sets the range: its high and low.
- Candle 2 trades beyond one side of that range (the sweep) and closes back inside it. If it sweeps both sides, skip it.
- Entry at candle 2’s close, toward the other side of the range: sell after a sweep of the high, buy after a sweep of the low.
- Stop just beyond candle 2’s wick. Target the other side of candle 1.
- Exit when candle 3 closes if neither is hit (we also tested holding for 3 candles).
Results on NQ
| Candles | Trades | Winners | Total, 1 NQ contract | Per trade |
|---|---|---|---|---|
| 1-hour | 15,454 | 47.5% | −$287K | −$19 |
| 4-hour | 3,383 | 49.3% | −$37K | −$11 |
| 4-hour, random direction | 3,383 | 49.7% | −$17K | −$5 |
| 4-hour, held for 3 candles | 2,917 | 47.8% | −$108K | −$37 |
On 4-hour candles the average win (+$695) and average loss (−$701) were almost the same size, and the worst drawdown was about −$105K. Only 4 of 10 years made money.
Every version, on newer data
We set the rules on older data (NQ before 2023, gold before 2014) and checked them on newer data. Results are in R, where 1R is the trade’s risk.
| Version | NQ, 2023–2026 | Gold, 2014–2020 |
|---|---|---|
| 1-hour candles | −0.05R | −0.31R |
| 4-hour candles | −0.05R | −0.13R |
| Daily candles | −0.07R | −0.05R |
| 4-hour, with the 20-day trend only | −0.04R | — |
| 4-hour, only the 1, 5 and 9 AM ET candles | −0.06R | — |
| 4-hour, target at least 1× the risk | −0.12R | — |
Verdict: myth. A sweep that closes back inside the range didn’t tell us which way price would go next, and costs did the rest. The setups that look perfect on a chart are the ones that worked; the losing ones look the same until they don’t.
Method and data
- NQ futures 1-minute data, April 2017 to September 9, 2026; 4-hour candles start at 1, 5 and 9 AM ET and so on. Gold: Oanda XAU/USD 1-minute data, 2006–2020.
- Trades are followed minute by minute; the stop is checked before the target when both fall in the same minute.
- Stop 0.05× the 14-candle ATR beyond the wick. One trade at a time. One contract ($20 per point), $15 per trade for costs on NQ; $0.35 per ounce on gold.
- Our fixed-rule version of a concept shared publicly; creators’ own rules and filters may differ. Hypothetical results.
Questions
Does Candle Range Theory work?
Not in our backtest. On NQ 4-hour candles from 2017 to 2026 it lost −$37K on 3,383 trades with one contract, and a random direction on the same trades lost less. Every timeframe we tried lost money on newer data, on NQ and gold.
What timeframe is best for CRT?
None of the ones we tested made money: 1-hour, 4-hour and daily candles all lost after costs on NQ futures and gold.
Does CRT work better with the 1 AM, 5 AM and 9 AM candles?
No. Using only those 4-hour candles on NQ still lost about −0.06R a trade on 2023–2026 data.
Keep reading
Cite this: Trading Decisions Lab, “Myth Test #17: Does Candle Range Theory work?” (Oct 31, 2026). https://www.tradingdecisionslab.com/research/candle-range-theory-backtest
Simulated backtest. Hypothetical or simulated performance results have certain limitations and do not represent actual trading; see the full disclaimer. Past results don’t guarantee future results. Educational only, not financial advice. Futures are leveraged and you can lose more than your deposit.