The stochastic oscillator is one of the most widely installed indicators on retail charts, and also one of the most misread. It does not predict reversals. It answers a narrow, useful question: where did price close relative to its recent high-low range? Once you hold that definition firmly, the indicator becomes a descriptive tool rather than a signal generator.
How it is built
Pick a lookback window, commonly 14 bars. Find the highest high and lowest low in that window. The raw reading, called %K, is:
%K = (latest close − lowest low) ÷ (highest high − lowest low) × 100
If the latest close equals the highest high of the window, %K is 100. If it equals the lowest low, %K is 0. A close in the middle of the range gives about 50. The second line, %D, is a short moving average of %K (typically 3 periods) and acts as a smoother. The "slow" version most platforms default to applies another smoothing step to %K before plotting.
Notice what is absent: volume, momentum in pips, and any view on fundamentals. The stochastic is a normalised position inside a range. Because it is normalised between 0 and 100, it looks comparable across pairs and timeframes, which is both its appeal and its trap.
Why "overbought" is a misleading word
Most platforms draw lines at 80 and 20 and label the zones overbought and oversold. In a range-bound market, a reading above 80 does mean price is near the top of the recent range, and fading it has historically worked more often than in trends. In a trending market, however, price can close near the top of its range for many consecutive bars. The oscillator will sit above 80 for as long as the trend persists, and anyone treating each visit as a sell trigger accumulates a string of small losses against the move.
A reading above 80 is better described as "price is closing near the top of its recent range." Whether that is a warning or a confirmation of strength depends on the market regime, not on the indicator.
Illustrative example
Suppose, purely for illustration, that over the last 14 daily bars a pair has a highest high of 1.1300 and a lowest low of 1.1100, a 200-pip range. Today it closes at 1.1280. %K = (1.1280 − 1.1100) ÷ (1.1300 − 1.1100) × 100 = 90. The oscillator says price closed in the top tenth of its range. It says nothing about what happens tomorrow.
Now imagine the same pair breaks out and the next day the window's highest high moves to 1.1350 while the low is still 1.1100. A close at 1.1340 gives %K = 96. The reading barely changed, yet the market made a new range high. This is why the stochastic is poor at distinguishing "stretched and about to snap back" from "strong and getting stronger."
Where it can help
- Identifying regime. If the oscillator keeps hugging the extremes for weeks, you are probably in a trend, and mean-reversion setups deserve scepticism. If it swings back and forth through the middle, the market is ranging.
- Timing within a decision you already made. If your plan says to buy pullbacks in an established uptrend, a stochastic that dips toward 20 and turns up shows the pullback has closed near the low of its range and is recovering. It refines entry timing; it should not create the idea.
- Spotting divergence, carefully. Price makes a higher high while the oscillator makes a lower high. This tells you the latest push closed less strongly within its range. It is a hint of fading momentum, not a guarantee, and divergences can persist for a long time.
Common mistakes
- Selling every cross above 80. Crossovers of %K and %D happen constantly. On a 5-minute chart you can see dozens a day, most of them noise.
- Optimising the settings. Switching from 14,3,3 to 9,5,2 after looking at a chart produces a pretty history that is curve-fitted. See our piece on backtesting honestly for how to test without fooling yourself.
- Ignoring the trend on a higher timeframe. A sell-looking reading on the 15-minute chart inside a strong daily uptrend is fighting the larger move. Our guide to multi-timeframe analysis covers this.
- Stacking similar indicators. Stochastic, RSI and MACD are all derived from the same price series. Three agreeing oscillators are one opinion repeated, not three confirmations. Compare with RSI and MACD to see how much they overlap.
A sensible way to use it
First classify the market: trending or ranging, using something simple such as the slope of a longer moving average or the structure of highs and lows. In a range, the stochastic helps you see when price is near an edge, and you can combine that with a support or resistance zone. In a trend, use it only to time pullback entries in the trend's direction, and keep your stop-loss based on market structure and volatility, never on the indicator itself.
Whatever you do, size every trade by the distance to your stop and a fixed risk percentage, as described in position sizing. An indicator can change your timing slightly. Risk control determines whether you survive long enough for timing to matter.
Takeaway
The stochastic oscillator tells you where the last close sits within a recent range. That is a fact, not a forecast. Treat 80 and 20 as descriptions of position, decide the regime first, and never let a single oscillator override your plan. This article is educational and illustrative; it is not a trade recommendation.