Stochastic Oscillator

The oscillator that measures where, not how fast

RSI measures momentum — how fast price is changing relative to itself. MACD measures the gap between two smoothed trends. ATR measures range magnitude. Stochastic does something different: it measures where in the recent range the close landed. That distinction, small on the surface, produces a differently-shaped indicator with differently-shaped signals.

George Lane (president of Investment Educators, Inc., Watseka, IL) popularized it. Murphy's attribution:

The Stochastic oscillator was popularized by George Lane. It is based on the observation that as prices increase, closing prices tend to be closer to the upper end of the price range. — John Murphy, Technical Analysis of the Financial Markets

Kaufman's one-sentence framing is the cleanest definition you'll get:

The stochastic indicator, created by George Lane, is an oscillator that measures the relative position of the closing price within a past high–low range.

Neither book gives a specific year, and neither mentions Ralph Dystant (sometimes co-credited in other sources). Treat the Lane/Dystant co-creation folklore as outside our reference library.

The formula

Raw %K — where in the last N bars' range did we just close:

At 100, the close equals the highest high of the last N bars. At 0, it's the lowest low. At 50, we're right in the middle. N is typically 14.

From raw %K, Lane layered two smoothings. Kaufman lays out the three-line system verbatim:

%K is the raw stochastic, without smoothing, %D is a 3-day smoothing of %K, and %D-slow is a 3-day smoothing of %D.

Which gives three flavors that get confusingly-named:

Murphy's own practical guidance:

Most traders use the slow stochastics because of its more reliable signals.

Raw %K is too whippy. The double smoothing that produces Slow Stochastic is the canonical trade-off: enough lag to filter noise, not so much that signals come late.

Default parameters: 14 for N (Murphy's explicit default) with 3/3 for K-smoothing / D-period — the widely-used "14/3/3 Slow" convention. Note: Murphy doesn't name this combo as a single "default preset"; he defaults to 14 and references the two 3-SMA smoothings separately. The 14/3/3 label is industry shorthand.

What "range position" actually tells you

Momentum (RSI) is about velocity. Stochastic is about location. They look similar on the chart but answer different questions:

• RSI = 70: "Price has moved up more than down, on average, over the last 14 bars." That's a statement about the path.
• %K = 70: "Today's close landed at the 70th percentile of the last 14 bars' range." That's a statement about the end-point.

Lane's underlying premise: in a healthy uptrend, closes cluster near the top of the range. Failure to close near the extreme is the first sign of a stall. That's a legitimately different insight than RSI provides — hence the common pairing.

Kaufman is explicit about the distinction:

This makes the stochastic oscillator conceptually different from the MACD, which uses the difference between two trends, and the RSI, which uses only the closing prices, and gives weight to the up and down moves.

Play with it

Toggle Fast / Slow / Full — watch the line get smoother and the signals get later. Flip between markets and notice:

• In a strong uptrend, %K and %D camp in the 80+ zone for weeks. Trading the "sell when Stochastic is above 80" rule mechanically here is a fast way to get run over.
• In sideways markets, %K and %D whip back and forth — useful for range-bound trading if you filter the noise.
• In downtrends, mirror: %K and %D pin the lower zone.

Kaufman's direct warning:

During a sustained downtrend, it is common for the indicator values to fall below the lower threshold and remain there for extended periods. Although the rolling calculation is intended to self-adjust to the full 0 to 100 range, a strong trend will prevent that from happening.

This is the same overbought/oversold trap we hit in the RSI lesson — only Stochastic is more sensitive, so the trap is deeper.

The signal hierarchy

Not all Stochastic signals are created equal. Murphy orders them:

1. Divergence is the primary signal. A bearish divergence: the %D line forms two declining peaks above 80 while price prints new highs. The bullish mirror: two rising bottoms below 20 while price makes new lows.
2. %K crosses %D is the trigger, not the signal. Murphy's explicit framing: "Assuming all of these factors are in place, the actual buy or sell signal is triggered when the faster K line crosses the slower D line."

The chain: divergence sets up the trade; the %K/%D crossover fires it. Mechanically trading every %K/%D cross without the divergence context produces noise, especially from Fast Stochastic.

Left vs right crossovers (Kaufman): the direction of %K relative to %D's turn matters:

• Right crossover — %D reverses first, then %K catches up and crosses. Kaufman: "a slow, stable change of direction and is a more favorable pattern."
• Left crossover — %K slams through %D at speed. More common, less reliable.

Lane's named patterns

Kaufman documents a handful of Lane-specific patterns that are worth knowing even if you don't trade them all:

• Hinge — "A reduction in the speed of either the %K-slow or %D-slow lines, shown as a flattening out, indicates a reversal on the next day." The most famous of Lane's patterns.
• Warning — "An extreme turn in the faster %K-slow (from 2 to 12%) indicates at most two days remaining in the old trend."
• Extremes — "Reaching the extreme %K-slow values of 0 and 100 requires seven consecutive days of closes at the highs (or lows). The test of these extremes, following a pullback, is an excellent entry point."
• Bear Setup — "Although the line chart shows higher highs and lows, if the %D-slow line has lower lows, a bear market setup has occurred." (This is a divergence with a different name.)
• Failure — %K-slow crosses %D-slow after penetrating an extreme, pulls back to %D-slow, but fails to re-cross. Often the strongest signal.

Lane was a serious pattern-classifier. Most modern users of Stochastic run it on autopilot without knowing these named reads exist.

Stochastic vs RSI

They look similar. They answer different questions. Murphy's one-paragraph comparison:

The RSI line is less volatile and reaches extremes less frequently than stochastics. The best signals occur when both oscillators are in overbought or oversold territory. — Murphy, on Figure 10.17's side-by-side

And the practical recommendation that gets skipped by most single-indicator users:

It's also a good idea to combine stochastics with RSI.

Kaufman's side-by-side observation (kaufman.txt:14960):

The stochastic (bottom panel) gives the overall impression that it moves faster and exaggerates the swings. There is no smoothing in the raw stochastic, so it has less lag than the other indicators.

The honest summary:

Neither is "better." They answer different questions; Murphy's pros use them together, taking trades only where both confirm.

What Bulkowski says — nothing

For completeness: Thomas Bulkowski, whose 150,000-sample database is the statistical standard for pattern testing, has no Stochastic content in Encyclopedia of Chart Patterns (3rd ed., 2021). No win rates, no failure-rate tables, no worked trade examples. Whatever statistical confidence you extend to Stochastic signals, it isn't coming from Bulkowski-style testing.

Kaufman's closest-to-a-test is a qualitative 2002 S&P futures walkthrough — a 20-day Stochastic with a 60-day MA trend filter caught two clean shorts that year. One case study is not a backtest. The prescriptive rule he extracts:

Enter a short sale on the first stochastic sell signal after the trend has turned down. Enter a long on the first stochastic buy signal after the trend has turned up.

Trend filter is carrying most of the weight in that rule. Stochastic is the trigger, not the edge.

Hidden traps

• Shorting strong uptrends because %K reads above 80. Same trap as RSI — amplified because Stochastic pins extremes faster. Strong trends can keep %K in the 80s for weeks.
• Trading every %K/%D cross. Without the divergence context (Murphy's primary signal), crossovers are noise. Especially on Fast Stochastic.
• Using raw %K. Kaufman: too volatile. Stick to Slow or Full.
• Treating overbought/oversold as trade triggers. These are warnings. The trade is confirmed by cross and divergence.
• Ignoring the higher timeframe. Murphy: "A daily signal is followed only when it agrees with the weekly signal. Two crossover systems in which this principle is especially true are MACD and Stochastics." Weekly dominates.
• Misreading the variant names. "Slow Stochastic" uses the second and third of three lines (%K-slow + %D-slow). "Fast Stochastic" uses the first and second (raw %K + %D). Reading the wrong lines off your charting platform is a depressingly common bug.

Quick check

What you now know

• Stochastic measures range position, not momentum — where today's close sits inside the last N bars' high-low range, on a 0–100 scale.
• George Lane popularized it (Investment Educators, Inc.). Neither of our reference books names a specific year or mentions Dystant.
• Three formula variants: Raw %K (too whippy), Slow Stochastic (canonical: 14/3/3), Full Stochastic (fully tunable).
• 80/20 are the standard thresholds — stricter than RSI's 70/30. Stochastic reaches extremes more often and deserves tighter cutoffs.
• Signal hierarchy: divergence in an extreme zone sets up the trade; %K/%D crossover confirms. Cross without divergence is noise.
• Right crossovers (%D turns first) are more reliable than left crossovers (%K slams through first).
• Lane's named patterns — Hinge, Warning, Extremes, Bear Setup, Failure — are worth knowing as refinements of the divergence-and-cross base.
• Stochastic is faster and noisier than RSI; pros use both and take signals where they agree.
• Higher-timeframe filter is non-negotiable. Murphy: for MACD and Stochastic especially, daily signals are only followed if the weekly agrees.
• Bulkowski has no Stochastic stats. Whatever statistical confidence you extend to the indicator, it's not coming from his testing framework.

Indicators Deep Dive unit complete. You now have the full oscillator + smoother + volatility toolkit: Moving Averages, RSI, MACD, Bollinger Bands, ATR, and Stochastic. Next up: Volume & Order Flow — stepping out of pure price-derived math and into volume analytics.