Narrow Range Days: The NR7 Idea and Volatility Cycles
A narrow range day is one where the daily range is smaller than in recent days. The best-known version, NR7, marks the day with the smallest range of the last seven days. The idea comes from the work of trader Toby Crabel.
The logic rests on volatility cycles. Periods of low volatility tend to be followed by higher volatility, because compressed ranges build up pressure for expansion. When the smallest range in seven days occurs, the following day or two have often shown larger moves, though the direction is unknown.
A simple way to use it is to place a buy stop above the NR7 day's high and a sell stop below its low for the following session, canceling one when the other triggers. The stop for each sits at the opposite side of the range, so the risk equals the narrow range, which is small by definition.
The strength of the idea is that risk is naturally tight. The weakness is false breakouts, as the market often probes one side and reverses.
Testing is simple. Mark every NR7 on your pair over several years, and record the next day's range relative to the average and whether a breakout followed through.
Results vary across pairs and eras. Treat published claims cautiously, and measure your own.
Practical step: find the NR7 days on your pair over the last year, and compute the average range of the next day compared with the overall average.
Narrow Range Days: The NR7 Idea and Volatility Cycles
Re: Narrow Range Days: The NR7 Idea and Volatility Cycles
Small detail that matters: an NR7 on a Friday or before a holiday is less interesting. Quiet days before quiet periods often just stay quiet. The same goes for the days before a big scheduled event, like a Fed decision, where the market is simply waiting. In that case the narrow range is about the calendar, not a volatility cycle building up pressure, and the expansion often happens on the event itself, which is a much harder trade. I'd add a filter when testing: exclude NR7 days followed by a holiday or with a major event the next day, and compare the results with and without them. The cleaner subset is usually more useful.
It’s Fairman 
Re: Narrow Range Days: The NR7 Idea and Volatility Cycles
The tight risk is the real appeal. Even with lots of false breaks, the losses are small, so it's worth measuring on your own pair rather than dismissing it. The thing to watch is cost. On a pair with a narrow daily range, the spread can be a meaningful part of the stop, especially if your orders trigger in the quiet Asian hours. You might get better results by only activating the orders from the London open, which for us is 8am or 9am depending on the season. Test both versions. Small details like activation time can turn a breakeven idea into a usable one, or show it doesn't work.
It’s Fairman 
-
FXS Prop Desk
- Posts: 1693
- Joined: Wed Oct 07, 2026 9:52 am
Re: Narrow Range Days: The NR7 Idea and Volatility Cycles
A practical note: NR7 on its own does not tell you direction, so many traders combine it with an inside day (NR7 that is also an inside day, sometimes called NR7-ID) and place orders on both sides of the range. The opposite order is cancelled once one side triggers.
Combining NR7 with a higher-timeframe trend filter can help decide which direction to favour when the breakout comes.
Combining NR7 with a higher-timeframe trend filter can help decide which direction to favour when the breakout comes.
FXS Official Team | The Idea Lab & Strategy
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Narrow Range Days: The NR7 Idea and Volatility Cycles
The size of the narrow day against a longer average seems to matter more than the label itself. Say the last seven EURUSD days ran 82, 95, 66, 88, 72, 56 and 31 pips, an average of 70, so the NR7 is under half of a normal day.
With entry orders 2 pips beyond each side the risk is 35 pips, and the chart (illustration) shows a next day of 94 pips, about three times the narrow range. Catching only half of a day like that covers the risk with room to spare.
Compare a week where every day sat between 40 and 48 pips and the NR7 printed at 40. It is technically the same signal, but nothing was compressed relative to the week, so the expected follow through is much weaker.
A filter such as an NR7 below 60 percent of the 20 day average range would separate those two cases. It cuts the number of signals a lot, which matters when a pair might only print a handful of strong ones in a quarter.
One edge case is the day after a false break. Does anyone keep the orders live for a second session when the first trigger reverses back inside, or treat the setup as finished?
With entry orders 2 pips beyond each side the risk is 35 pips, and the chart (illustration) shows a next day of 94 pips, about three times the narrow range. Catching only half of a day like that covers the risk with room to spare.
Compare a week where every day sat between 40 and 48 pips and the NR7 printed at 40. It is technically the same signal, but nothing was compressed relative to the week, so the expected follow through is much weaker.
A filter such as an NR7 below 60 percent of the 20 day average range would separate those two cases. It cuts the number of signals a lot, which matters when a pair might only print a handful of strong ones in a quarter.
One edge case is the day after a false break. Does anyone keep the orders live for a second session when the first trigger reverses back inside, or treat the setup as finished?
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Narrow Range Days: The NR7 Idea and Volatility Cycles
The opposite side of the range is not the only place for the stop, and a tighter one is worth testing. A stop at the midpoint of the narrow day halves the risk, at the cost of being knocked out by more of the ordinary probes.
Say the NR7 runs from 1.0850 to 1.0880, with a buy stop at 1.0882. The far side stop at 1.0848 risks 34 pips, while a midpoint stop at 1.0865 risks 17.
With a 50 pip target, suppose the far stop wins 40 percent of the time and the midpoint stop only 30 percent. The first works out to 20 minus 20.4, a small loss of 0.4 pips per trade, while the second gives 15 minus 11.9, a gain of 3.1 pips.
A time limit on the orders matters too. If neither side triggers by the end of the following day, the compression has often turned into a longer squeeze, and any new NR7 is a fresh setup rather than a continuation of the old one.
Does anyone treat two NR7 days in a row as a stronger signal, or simply move the orders to the newer range? With the far side stop, the narrower second day also shrinks the risk again.
Say the NR7 runs from 1.0850 to 1.0880, with a buy stop at 1.0882. The far side stop at 1.0848 risks 34 pips, while a midpoint stop at 1.0865 risks 17.
With a 50 pip target, suppose the far stop wins 40 percent of the time and the midpoint stop only 30 percent. The first works out to 20 minus 20.4, a small loss of 0.4 pips per trade, while the second gives 15 minus 11.9, a gain of 3.1 pips.
A time limit on the orders matters too. If neither side triggers by the end of the following day, the compression has often turned into a longer squeeze, and any new NR7 is a fresh setup rather than a continuation of the old one.
Does anyone treat two NR7 days in a row as a stronger signal, or simply move the orders to the newer range? With the far side stop, the narrower second day also shrinks the risk again.
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Narrow Range Days: The NR7 Idea and Volatility Cycles
Most strategies are found by staring at charts until a pattern appears. That approach has a high false-positive rate: the human eye finds patterns in noise, and the one that "works" in the sample you looked at is probably luck.
An alternative starts with a hypothesis, a reason why an edge might exist, then tests it. For example: "Liquidity is thin in the minutes after the daily rollover, so price overshoots and mean-reverts." That statement specifies a mechanism, a time window and a direction. It can be tested on data the idea wasn't derived from.
Good hypotheses draw on market structure: who is trading, why, under what constraints, and how that creates predictable behaviour. Fixings, option expiries, rebalancing flows, central bank communication cycles and liquidity patterns all give reasons for something systematic.
Testing then asks a narrow question: does the predicted behaviour appear, and is it large enough to cover costs? If not, the hypothesis is rejected, and a clean negative result is still progress.
A written research log, with each hypothesis, test and outcome, prevents re-testing dead ideas and guards against selective memory.
Examples of hypotheses worth testing: "Spreads are cheapest in the first hour of the London session, so cost-sensitive strategies should concentrate there." "After large overnight gaps in yen pairs, intraday ranges are above average." "The first move after a Fed speaker headline is partially reversed within ten minutes." Each is specific, falsifiable and tied to a plausible mechanism.
A habit that helps is writing the expected result before running the test. If the outcome contradicts the expectation, that's information. If it matches, it's worth checking that the test wasn't accidentally built to confirm the idea.
Over time, a research log with dozens of entries becomes one of the most valuable assets a trader owns, more so than any single strategy.
What hypotheses have been tested lately, and which one was rejected most convincingly?
An alternative starts with a hypothesis, a reason why an edge might exist, then tests it. For example: "Liquidity is thin in the minutes after the daily rollover, so price overshoots and mean-reverts." That statement specifies a mechanism, a time window and a direction. It can be tested on data the idea wasn't derived from.
Good hypotheses draw on market structure: who is trading, why, under what constraints, and how that creates predictable behaviour. Fixings, option expiries, rebalancing flows, central bank communication cycles and liquidity patterns all give reasons for something systematic.
Testing then asks a narrow question: does the predicted behaviour appear, and is it large enough to cover costs? If not, the hypothesis is rejected, and a clean negative result is still progress.
A written research log, with each hypothesis, test and outcome, prevents re-testing dead ideas and guards against selective memory.
Examples of hypotheses worth testing: "Spreads are cheapest in the first hour of the London session, so cost-sensitive strategies should concentrate there." "After large overnight gaps in yen pairs, intraday ranges are above average." "The first move after a Fed speaker headline is partially reversed within ten minutes." Each is specific, falsifiable and tied to a plausible mechanism.
A habit that helps is writing the expected result before running the test. If the outcome contradicts the expectation, that's information. If it matches, it's worth checking that the test wasn't accidentally built to confirm the idea.
Over time, a research log with dozens of entries becomes one of the most valuable assets a trader owns, more so than any single strategy.
What hypotheses have been tested lately, and which one was rejected most convincingly?