Is the Stochastic RSI better than RSI?

Neither dominates; they answer different questions. RSI measures momentum directly and suits regime and divergence work; the Stochastic RSI measures where that momentum sits within its recent range and suits timing. The extra sensitivity produces more signals and more false ones, so the choice depends on holding period and the filters around it.

What do 80 and 20 mean on the Stochastic RSI?

That RSI is near the top or bottom of its own lookback range, not that price reached any absolute level. Because the scaling is min-max, those zones are hit far more often than RSI's 70/30, sometimes several times within one price swing. Many traders therefore treat the exit from the zone, not the entry, as the event.

What are the standard Stochastic RSI settings?

The widespread default is 14 for the RSI, 14 for the stochastic window, and 3,3 for %K/%D smoothing. Shortening either window makes an already fast tool faster; lengthening the window or the smoothing calms it down. There is no best set, and parameters tuned to one instrument rarely transfer unchanged to another.

Why does the Stochastic RSI hit 0 and 100 so often?

Min-max mechanics. The formula divides RSI's position by the width of its recent range, so any new lookback high in RSI prints exactly 100 and any new low prints 0, even if RSI itself moved only a few points. A pinned reading means RSI is at a local extreme, not that a reversal is due.

Can I trade the Stochastic RSI on its own?

It is rarely used standalone. Two derivations removed from price, it says nothing about structure, levels, or trend, and its crosses fire constantly in chop. Typical usage subordinates it to a directional filter (trend, structure, or a higher timeframe) and reserves it for the timing decision inside that context.

Stochastic RSI

The Stochastic RSI applies the stochastic formula to RSI values instead of prices. It is a Momentum & Oscillators concept. The Library holds 1 implementation, ready to run on Vela.

The standard Stochastic RSI indicator

Stochastic RSI exactly as classically defined: the faithful reference build of the original formula, free to run in Quant.

Stochastic RSI

Indicator

What is the Stochastic RSI?

The Stochastic RSI applies the stochastic formula to RSI values instead of prices. Introduced by Tushar Chande and Stanley Kroll in The New Technical Trader (1994), it measures where the current RSI sits within its own recent range: current RSI minus the lowest RSI of the lookback, divided by the highest RSI minus the lowest RSI. The result runs from 0 to 1, commonly rescaled to 0-100, with 14 the usual length for both the RSI and the stochastic window.

The motivation was RSI's habit of trending without touching its extremes: in a sustained move RSI can hold mid-scale for long stretches and never trigger 70/30 rules. Re-scaling RSI against its own range fixes that mechanically, because the Stochastic Oscillator arithmetic prints 100 whenever RSI makes a lookback high and 0 whenever it makes a lookback low. Platforms typically smooth the raw line with a 3-period %K average and a 3-period %D signal.

The trade-off is the whole story: the Stochastic RSI reaches overbought/oversold territory far more often than RSI itself, which makes it useful for timing and noisy as a standalone signal. It is an indicator of an indicator, two derivations removed from price, so it can pin at a boundary on a modest move. Many workflows subordinate it to a trend filter or structure rather than trading its crosses in isolation.

How to calculate the Stochastic RSI

The calculation runs a stochastic over an RSI series instead of over price.

1Compute RSI on the chosen source, typically over 14 periods.
2Across the most recent N RSI values (14 is the common default), record the highest and lowest readings.
3Divide (current RSI minus the lowest RSI) by (highest RSI minus the lowest RSI). Multiply by 100 for the 0-100 display; by construction the value touches a boundary whenever RSI makes an N-period extreme.
4Smooth for reading: %K is commonly a 3-period simple average of the raw line, %D a 3-period average of %K, with 80/20 (or 0.8/0.2) as the conventional zones.

How it's calculated

A stochastic oscillator applied to RSI values, measuring where RSI sits inside its own recent range.

\mathrm{LL}_t = \min(\operatorname{RSI}) \text{ over the last } m \text{ bars}

\mathrm{HH}_t = \max(\operatorname{RSI}) \text{ over the last } m \text{ bars}

\operatorname{StochRSI}_t = 100 \times \frac{\operatorname{RSI}_t - \mathrm{LL}_t}{\mathrm{HH}_t - \mathrm{LL}_t}

\%K_t = \operatorname{SMA}(\operatorname{StochRSI}_t, k)

\%D_t = \operatorname{SMA}(\%K_t, d)

t: current bar index

RSI: Wilder relative strength index of closes, length n

n: RSI length (commonly 14)

RSI_t: RSI value at the current bar

m: stochastic lookback applied to the RSI series (commonly 14)

LL_t: lowest RSI value of the last m bars

HH_t: highest RSI value of the last m bars

StochRSI_t: stochastic of RSI, scaled to 0..100

SMA(x, len): simple moving average of series x over len bars

k: %K smoothing length (commonly 3)

%K_t: smoothed StochRSI line

d: %D smoothing length (commonly 3)

%D_t: signal line, an SMA of %K

Chande and Kroll's original formula is unscaled between 0 and 1; platforms multiply by 100 and add the %K and %D smoothing.

If RSI is flat over the window, HH_t equals LL_t and the ratio is undefined; implementations output 0 or hold the prior value.

As a stochastic of an oscillator it tags 0 and 100 far more often than RSI reaches its own extremes.

How traders use it

As a pullback timer inside trends: with a higher-timeframe trend filter defining direction, an oversold Stochastic RSI hooking back up times entries with the trend while skipping counter-trend fades.
As a fast mean-reversion gauge in ranges, where its frequent trips to the extremes match the cadence of range rotation better than the slower parent RSI does.
As a crossover trigger: %K crossing %D inherits the stochastic's grammar, and many workflows only honor crosses that occur inside the extreme zones to filter mid-scale noise.
As a normalized component in composites: because it maps RSI onto a fixed 0 to 1 range, it slots cleanly into scoring systems and dashboards that need comparable scales across inputs.

Stochastic RSI vs related tools

RSI: RSI is the raw material: smoothed strength of up closes versus down closes, bounded 0-100 but slow to reach its extremes. The Stochastic RSI re-scales those values against their own recent range, so it cycles hard between 0 and 100 while RSI drifts mid-band.

Stochastic Oscillator: Identical formula, different input: the stochastic locates the close within the price range, the Stochastic RSI locates RSI within the RSI range. One derivation from price versus two, which is why the Stochastic RSI is faster and jumpier.

RSI-2: RSI-2 speeds RSI up by shrinking the lookback to two periods instead of re-scaling: same arithmetic, hair-trigger length. The Stochastic RSI keeps the standard RSI and changes the lens through which it is read.

Connors RSI: Connors RSI blends three components (a short RSI of price, an RSI of the up/down streak, and a percent rank of returns) into one composite. Both aim at faster mean-reversion timing than classic RSI, one by composition, the other by range-scaling.

Stochastic RSI FAQ

Turn Stochastic RSI into a trading strategy.

Open the implementation from this page in Vela. Quant, our coding agent, builds the strategy with you and backtests it on real data.

Build a strategy in Vela

Browse this familyMomentum & Oscillators85 concepts mapped · 85 in the Library

Previous conceptStochastic Pop Next conceptSwing Index