A stock breaks above a two-week high on what looks like clean volume. You enter. Within three bars it reverses back inside the range and stops you out. This happens because range expansion alone tells you nothing about the quality of the expansion. The Range Expansion Index (REI) was built to solve exactly this problem. It measures whether a price move has the internal structure of a genuine directional push or just the appearance of one.
I use the REI as a secondary filter specifically on breakout setups. When Donchian Channel Width expands or a key level breaks, the REI tells me whether the move has conviction behind it or whether price is just poking its head out before reversing. It does not replace your entry signal. It tells you whether to trust it.
What the Range Expansion Index Measures
Thomas DeMark introduced the REI as part of his broader work on market timing. The indicator compares the current bar’s high and low to the highs and lows from two bars ago. It then applies a set of conditional logic tests to determine whether the price movement qualifies as genuine expansion or noise.
The output oscillates between -100 and +100. Readings above +45 indicate overbought conditions where upside expansion may be exhausting. Readings below -45 indicate oversold conditions. Between -45 and +45, the market is considered trendless or consolidating.
What makes the REI different from a standard oscillator: it does not just measure speed or magnitude. It tests the geometric relationship between consecutive bars. A large candle that fails specific high-low comparison tests will not register as expansion. This is the filter most traders miss. A big bar is not the same as a directional bar.
How the REI Calculation Works
The REI calculation has two layers. First, each bar is tested against conditions that define whether a move qualifies as range expansion. Second, the qualifying moves are summed and normalized.
For each bar, DeMark compares the current high and low to the high and low from two periods earlier. Two conditions must be met for a move to count as expansion:
Condition 1: The current high is greater than or equal to the low from two bars ago, OR the high from two bars ago is greater than or equal to the current low.
Condition 2: The current high minus the high from two bars ago has the same sign (both positive or both negative) as the current low minus the low from two bars ago.
When both conditions are met, the numerator value for that bar equals:
\text{Numerator} = (\text{High} - \text{High}_{-2}) + (\text{Low} - \text{Low}_{-2})
When conditions are not met, the numerator is zero. The denominator sums the absolute values of those same differences over the lookback period regardless of conditions:
\text{Denominator} = \sum_{i=1}^{n} \left( |\text{High}_i - \text{High}_{i-2}| + |\text{Low}_i - \text{Low}_{i-2}| \right)
The REI formula is then:
\text{REI} = \frac{\sum \text{Numerator}}{\text{Denominator}} \times 100
The default lookback is 5 periods. I find 5 works well for daily charts. On intraday timeframes, some traders extend to 8 to smooth out noise, but I have not found a compelling reason to deviate from the default for swing setups.
The Zero Line Trap
Most traders who first encounter the REI focus on the +45 and -45 levels. That is the wrong starting point. The real information sits around zero.
When the REI hovers near zero for several consecutive bars, it means the conditional tests are failing. Price is moving, but the moves lack the geometric structure DeMark’s logic requires. This is the indicator telling you that what looks like directionality on a candlestick chart is actually random chop.
I have seen this pattern repeatedly on daily charts around earnings season. A stock drifts higher for three or four days on light volume. The chart looks bullish. But the REI stays pinned between -10 and +10 because the high-to-high and low-to-low comparisons are not aligning directionally. Then the stock gaps down on the report and everyone wonders what happened. The REI had already flagged the move as structurally weak.
The common mistake: treating REI near zero as “no signal.” It is a signal. It is telling you to stay out.
Confirming Breakouts With the REI
The strongest use case for the REI is as a breakout confirmation filter. Pair it with any level-based breakout system and you reduce false entries significantly.
The setup works like this. Price breaks above resistance or a channel high. Before entering, check the REI. If the REI is above +45, the move has the internal structure of genuine expansion. If it is between 0 and +45, the breakout lacks conviction and you wait. If it is near zero or negative, you skip the trade.
Take AAPL on April 30, 2026. The stock opened at $270.50, pushed to a high of $276.00, pulled back to a low of $268.14, and closed at $271.35. Then on May 1, it gapped up to $278.86 and ran to $287.22 with a close of $280.14. That second day showed clean range expansion: the high-to-high comparison from two bars earlier ($287.22 versus $271.04 on April 29) was strongly positive, and the low-to-low comparison ($278.37 versus $267.04) aligned in the same direction. Both conditions met. REI would register a strong positive reading on that bar.
Compare that to MSFT over the same window. On April 30, MSFT dropped from $410.81 to close at $407.78 with a low of $398.01. On May 1, it bounced to a high of $417.11 and closed at $414.44. Looks like a reversal bounce. But the high-to-high comparison ($417.11 versus $426.82 on April 29) is negative while the low-to-low ($410.44 versus $420.29) is also negative. The conditions align, but in the bearish direction. REI would not confirm a bullish breakout here because the two-bar structural comparison still points down.
This is where the REI adds value that a simple volume confirmation check cannot. Volume tells you participation. REI tells you directional structure.
REI Versus Other Volatility and Momentum Filters
The REI occupies a niche that overlaps with several indicators but matches none of them exactly.
Bollinger Band Width measures whether volatility is compressed or expanded. It tells you a move might be coming but says nothing about direction or quality. The REI tells you whether a move that already started has the right structure.
The Relative Volatility Index measures standard deviation in directional terms. It is smoother and better for trend confirmation over longer windows. The REI is faster and more granular. It reacts within 2-3 bars of a structural change.
Historical Volatility tells you the realized range of recent moves. It is backward-looking by definition. The REI’s conditional logic is forward-looking in the sense that it tests whether current structure supports continuation.
The wrong approach: using REI as a standalone oscillator like RSI or Stochastics. It was not designed for that. Its strength is its conditional logic. Use it to confirm or deny setups identified by other tools.
False Signals and Where the REI Breaks Down
The REI is not immune to failure. Three situations produce misleading readings consistently.
First, gap-driven moves. When a stock gaps up 5% on news, the high-to-high and low-to-low comparisons are both strongly positive. REI goes to extreme readings. But the move has already happened. Entering on an REI reading of +80 after a gap is chasing, not confirming. I only use the REI on breakouts that develop intraday or over two sessions, not on gap openings.
Second, low-liquidity names. The conditional tests assume that high and low prices reflect real supply and demand. In thinly traded stocks, a single order can push the high well above where real interest exists. The REI will register expansion that is just a wide spread, not genuine demand.
Third, the lookback is short. Five bars means the REI resets quickly. After a strong expansion move, the REI can drop back near zero within three bars even if the trend is intact. Do not read a falling REI as a reversal signal. It just means the current rate of expansion is slowing relative to the prior surge. The trend can continue at a lower velocity and the REI will be near zero.
Practical Setup: REI Plus Donchian Channel Width
I pair the REI with Donchian Channel Width for a two-filter breakout system. The logic is simple.
Donchian Channel Width identifies when a range is expanding. That is the first condition: the channel is widening from a recent squeeze. But width expansion alone does not tell you whether the breakout bar has directional commitment. That is where the REI steps in.
Entry rules: Donchian Channel Width rising from a 20-day low. Price closes above the upper Donchian Channel. REI above +45 on the breakout bar.
Exit rules: REI drops below zero on a closing basis. This does not mean the trend is reversing. It means the expansion structure has broken. Tighten your stop or take partial profits.
Using SPY as context: on April 30, 2026, SPY pushed from a low of $710.45 to a high of $719.79, closing at $718.66. The prior two bars (April 28 close $711.69, April 29 close $711.58) showed a flat range. The April 30 bar broke above this congestion. The high-to-high comparison ($719.79 versus $712.20 on April 29) was strongly positive, and the low-to-low ($710.45 versus $709.25) moved in the same direction. REI would confirm. The next day, SPY continued to $724.87. That is the kind of setup the filter catches.
Settings and Timeframe Considerations
The default 5-period lookback works on daily charts for most liquid instruments. On weekly charts, the indicator becomes too slow to be useful for entry timing, though it can serve as a trend-quality overlay.
On 15-minute and hourly charts, the REI works but the conditional tests fire more often because intraday bars tend to have more random high-low variation. Expect more signals near the +45/-45 thresholds and more whipsaws. I increase the threshold to +55/-55 on intraday charts to compensate. This is not a published best practice. It is what I found works after watching the indicator on 15-minute S&P futures charts where the default thresholds triggered too frequently during midday chop.
Do not optimize the lookback period to fit recent price action. DeMark designed the 5-period setting around the conditional logic. Changing it to 3 makes the indicator too reactive. Changing it to 10 smooths out the conditional tests so much that you lose the binary signal quality that makes the REI useful in the first place.
When the REI Tells You to Walk Away
The most profitable use of the REI is the trades it keeps you out of. A breakout that lacks REI confirmation is not necessarily a short. It is a skip. The distinction matters.
When the REI stays near zero during a breakout attempt, the market is testing a level without committing. Sometimes it will push through anyway, and you will miss the move. That is the cost of the filter. But the moves that fail when the REI is near zero tend to fail hard. They reverse back into the range within one to three bars and stop out breakout traders who entered without confirmation.
I keep a simple rule: if the REI does not confirm the breakout bar, I do not enter on that bar. I can enter on a subsequent bar if the REI catches up. But the initial breakout bar with a flat REI is not worth the risk.
The REI is a narrow tool. It does one thing well: test whether a price move has the structural geometry of a directional push. Use it for that. Do not try to build a trading system around it alone. Combine it with channel tools, volume confirmation, and proper risk management for a filter that reduces false breakout losses without overcomplicating your process.
Educational content only. Not investment advice. Trading involves risk. You are responsible for your decisions.