The Slow Stochastic’s %K Is the Fast Stochastic’s %D — the Same Label Names a Different Line in Each Version

This page previously said %D is where implementations diverge. Source 3 shows exactly how. “There are three versions of the Stochastic Oscillator available on SharpCharts.” “The Fast Stochastic Oscillator is based on George Lane's original formulas for %K and %D.” “In this fast version, %K can appear rather choppy.”^[3]

The slow version is not a different formula — it is the fast version shifted one place along: “The Slow Stochastic Oscillator smooths %K with a 3-day SMA, which is exactly what %D is in the Fast Stochastic Oscillator.” And therefore — “Notice that %K in the Slow Stochastic Oscillator equals %D in the Fast Stochastic Oscillator”^[3]

So “%K” does not name one series. In the fast version it is the raw ratio; in the slow version it is that ratio already smoothed. Two charts labelled %K can be two different lines, which is what makes settings comparisons between platforms unreliable unless the version is stated.^[3]

“The Full Stochastic Oscillator is a fully customizable version of the Slow Stochastic Oscillator.” The defaults source 3 uses: “The default parameters were used in these examples: Fast Stochastic Oscillator (14,3), Slow Stochastic Oscillator (14,3) and Full Stochastic Oscillator (14,3,3).”^[3]

Source 3 quotes Lane on what the indicator tracks: it “doesn't follow price, it doesn't follow volume or anything like that. It follows the speed or the momentum of price. As a rule, the momentum changes direction before price.”^[3]

And on what may be acted on, it reports a far narrower claim than the usual overbought/oversold reading: “This was the first and most important signal that Lane identified.” — “Lane asserts that a %D divergence is the “only signal which will cause you to buy or sell.””^[3]

That is reported as Lane's assertion, by source 3. This page neither endorses it nor treats it as established — Lane's own writing was not consulted.^[3]

Source 3 also records a second construction of his — “George Lane identified another form of divergence to predict bottoms or tops, dubbed “set-ups.”” — and immediately qualifies it: “Remember that this is a set-up, not a signal.”^[3]

“The oscillator ranges from zero to 100. No matter how fast a security advances or declines, the Stochastic Oscillator will always fluctuate within this range.” “Traditional settings use 80 as the overbought threshold and 20 as the oversold threshold.” With the worked example: “Assume that the highest high equals 110, the lowest low equals 100, and the close equals 108.” — giving 80.^[3]

But the extremes are not signals by themselves: “Overbought readings aren't necessarily bearish. Securities can become overbought and remain overbought during a strong uptrend.” “In a similar vein, oversold readings aren't necessarily bullish. Securities can also become oversold and remain oversold during a strong downtrend.”^[3]

Its practical advice is directional rather than symmetric: “Look for occasional oversold readings in an uptrend and ignore frequent overbought readings.”^[3]

And it says plainly that the common triggers misfire: “Signal line crosses, moves below 80, and moves above 20 are frequent and prone to whipsaw.” “As KSS shows, early signals are not always clean and simple.”^[3]

The formula is %K = 100 x (Price - Low_N) / (High_N - Low_N), and the source defines the bounds plainly: "High and Low are the highest and lowest prices in the last 5 days respectively" in the five-period example. The window is the last N periods, which includes the period being evaluated.^[1]

This matters for anyone testing the indicator. With some indicators, using the current bar's high and low to compute that bar's value is a mistake — the value ends up describing the bar it is being compared against. Here it is not a mistake: the definition places the current bar inside the range. A %K that excluded the current bar would be a different indicator. This page does not survey how platforms implement it, only what the definition says.^[1]

A related structural point: the window is fixed. Because %K reads the highest and lowest prices of N periods rather than an exponentially weighted average, the value does not depend on how much history preceded the calculation. This is a contrast drawn by this page, not a statement in the source, and it distinguishes the stochastic from indicators built on smoothed averages.^[1]

%D is "the N-day moving average of %K (the last N values of %K)" — a smoothing of the raw line. And the source immediately flags the ambiguity: "Usually this is a simple moving average, but can be an exponential moving average for a less standardized weighting for more recent values." So two platforms can both be correct and still disagree — a comparison should state which one it used.^[1]

And the source restricts what %D alone can be used for. "There is only one valid signal in working with %D alone — a divergence between %D and the analyzed security." By the reference's own account, reading %D on its own for anything other than divergence is outside what it endorses. The usual construction pairs the lines: an alert exists "when the %D line is in an extreme area and diverging from the price action", and "the actual signal takes place when the faster %K line crosses the %D line."^[1]

Parameters are conventions, as elsewhere in this family. "Typical values for N are 5, 9, or 14 periods" and "smoothing the indicator over 3 periods is standard." The article also describes a three-line variant giving "an anticipatory signal in %K, a signal in the turnaround of %D at or before a bottom, and a confirmation of the turnaround in %D-Slow." Three named lines, three different roles — answers that treat "the stochastic" as one line lose that.^[1]

The article does not hold one definition of %D throughout. Its definition section gives "%D is the N-day moving average of %K (the last N values of %K)", but its worked calculation gives %D as a 3-period simple moving average, SMA_3(%K), with %D-Slow a 3-period average of %D. N-period and 3-period are not the same rule, and both appear in the same article. The summary line of this page follows the definition section; anyone reproducing the numbers should say which of the two they used. This gap is a reading of the source by this page, not something the source points out about itself.^[1]

The reasoning behind the indicator is written out. "The idea behind this indicator is that prices tend to close near the extremes of the recent range before turning points." The price is then "expressed as a percentage of this range with 0% indicating the bottom of the range and 100% indicating the upper limits of the range over the time period covered."^[1]

"Tend to" is the source's wording and is kept here. The article presents this as the premise the indicator is built on, not as a demonstrated regularity, and it offers no test of it. Whether acting on the resulting signals produces a profit is not addressed by this page, which carries no backtest figures.^[1]

One further reading is named. A "stochastic pop" occurs "when prices break out and keep going", interpreted as a signal to increase the current position, or liquidate if the direction is against the current position. That is the opposite response to the mean-reversion reading of an extreme %K — the same indicator supports both, and the article does not choose between them.^[1]

This page rests on two encyclopedia articles and one charting platform's own documentation. Lane's own writing and the interview quoted by source 3 were not consulted directly.^[1][2][3]

Vendor limitation: source 3 is StockCharts describing the versions it offers on its own charts. Its three-version scheme and its default parameters are that platform's arrangement, and another package may name or combine them differently — which is the practical reason the label ambiguity above matters.^[3]

Not established: whether a %D divergence really is the only actionable signal, as Lane is quoted saying; whether any stochastic rule is profitable. No backtest figures of this page's own appear anywhere on it.^[1][2][3]

Retrieved 2026-09-28. Source 3 HTTP 200, read in full (207 lines); 28 quotations from it were each searched for in the fetched text before publication.^[3]