Trend & directionRandom Walk Index · RWI
Michael Poulos's measure of how far price movement departs from what pure chance would produce.
Works in most conditionsEngine-computed on a fixed sample series
What it is
The Random Walk Index, created by Michael Poulos, measures how far a market's price movement departs from what pure chance — a random walk — would be expected to produce, in order to distinguish a genuine trend from directionless noise. It answers a foundational question that most indicators skip: is this market actually trending, or is its movement statistically indistinguishable from randomness? Poulos reasoned that in a random walk, price wanders and the distance it covers grows only with the square root of time, so any move that travels much farther than that in a straight line must reflect a real, non-random trend. The RWI compares the actual distance price has traveled to the distance a random walk of the same volatility would be expected to cover. For a beginner it is best understood as a statistical trend detector that asks whether a move is too orderly to be an accident.
How it is calculated
For a given lookback of n bars, the RWI of the highs is the current high minus the low of n bars ago, divided by the expected random-walk distance — the average true range times the square root of n. The RWI of the lows is the high of n bars ago minus the current low, over the same denominator. The square-root-of-n term is the crux, because it encodes the statistical fact that a random walk's expected displacement grows with the square root of the number of steps, so the denominator represents how far chance alone should carry price. The calculation is typically performed across a range of lookback values and the maximum is taken, capturing trends of different lengths. A ratio well above 1 means the actual move dwarfs the random expectation, a real trend, while a value near or below 1 means the movement is within the bounds of noise.
Reading it, step by step
A high RWI of the highs points to a genuine, statistically significant uptrend, and a high RWI of the lows to a genuine downtrend, whereas both readings sitting low marks a directionless, essentially random market. The indicator is fundamentally about separating trend from chop rather than giving buy and sell triggers directly. A reading above 1 is the threshold where movement begins to exceed random expectation, and the further above 1, the more orderly and significant the trend. When the RWI of the highs is elevated while the RWI of the lows is subdued, the market is trending up; the reverse indicates a downtrend; both low means stand aside. Rising RWI in one direction signals a strengthening trend, while a collapse of both toward random levels signals the trend has dissolved into noise.
Best timeframes and settings
The RWI can be applied across intraday, daily, and weekly charts, and its lookback range should be matched to the horizon being traded — shorter lookbacks for intraday trend detection, longer ones for swing and position work. The calculation and its lookback are sensitive, and there is no universal threshold that cleanly separates trend from noise across all instruments, so the levels often need calibration per market. A shorter lookback reacts faster to emerging trends but is noisier, while a longer one is steadier but slower to recognize a turn. Because it is computed over a range of lookbacks, the choice of that range shapes which trend lengths the indicator is tuned to detect. Traders typically settle on a lookback band suited to their timeframe and use the RWI as a filter rather than fine-tuning it constantly.
When and where to use it
The RWI works best as a regime filter, layered under an entry system to answer whether a trend is present before trend-following signals are trusted, and it applies across equities, futures, and forex. Its job is precisely to identify the regime — trade trend-following tactics when it confirms non-random movement in your direction, and switch to range tactics or stand aside when it collapses toward random levels. It is less useful as a standalone entry signal, since it tells you a trend exists but not the precise moment to act. It is also less widely supported in charting platforms than ADX, which serves a similar purpose, so availability can be a practical consideration. Reach for it when you want a statistically grounded way to decide whether the market is trending at all.
Strategies that use it
The core strategy uses the RWI as a filter: only take long trend-following entries when the RWI of the highs is meaningfully above 1 and confirming an uptrend, only take shorts when the RWI of the lows confirms a downtrend, and stand aside or switch to mean-reversion tactics when both are near random levels. A regime-switching approach runs two playbooks — trend-following when the RWI is high and range-fading when it is low — using the indicator to toggle between them. A confirmation approach pairs the RWI read with a separate entry trigger, such as a breakout or moving-average cross, taking the trigger only when the RWI validates that the move is non-random. In each case the RWI governs whether to trade the trend while a separate tool supplies the exact entry, since the RWI is a regime detector rather than a timing signal.
Combining it with other indicators
The Average Directional Index is the RWI's closest analogue, and the two can cross-check each other's read on whether a trend exists and how strong it is. The Vortex Indicator adds a directional trend read that complements the RWI's trend-versus-noise measure. A linear-regression slope quantifies the direction and steepness of the trend the RWI has confirmed is real. Moving averages or breakout levels supply the entry triggers that the RWI filters. The consistent logic is to use the RWI to establish that a trend is statistically present and then let a directional or entry tool act within that confirmed regime, rather than relying on the RWI for timing.
Where it fails
The RWI's calculation and lookback are sensitive, and no single threshold cleanly separates trend from noise across all instruments, so a level that works on one market may misfire on another and requires calibration. It is a regime detector, not an entry signal, so traders who expect precise buy and sell timing from it will be disappointed. It is also less widely implemented than ADX, and its statistical framing, which assumes a particular random-walk model of volatility, can misjudge markets that gap or behave non-normally. The remedies are to calibrate thresholds per instrument, to use it strictly as a filter beneath a separate entry system, and to cross-check it against a more common trend gauge like ADX. Treating a single RWI reading as a mechanical trigger rather than as context is the central mistake.
A worked example
Suppose over a 14-bar lookback a stock's current high is 60, the low 14 bars ago was 50, its average true range is about 1.20, and the square root of 14 is roughly 3.74, so the expected random-walk distance is about 1.20 times 3.74, or 4.49 points. The actual upward travel of 10 points divided by that 4.49 gives an RWI of the highs near 2.2, well above 1, signaling that the move is far too orderly to be random and that a genuine uptrend is in force. Meanwhile the RWI of the lows is near 0.6, subdued, confirming the trend is one-directional. A trend trader takes this as validation to trust long entries from a separate trigger and to avoid fading the move. Weeks later, as price chops sideways, both the RWI of the highs and of the lows fall toward 0.8, near random levels, telling the trader the trend has dissolved into noise and that range tactics, or standing aside, are now appropriate.