Stochastic Oscillator Explained: Settings, Signals and Limits

Your chart says the stochastic is at 92, “overbought”, and the pair keeps climbing for another two weeks. The indicator did exactly what it was built to do. The problem was the meaning you gave the number.

Quick answer: The stochastic oscillator measures where the latest close sits within the high-low range of the last 14 candles, on a scale from 0 to 100. A reading of 75 means the close was three-quarters of the way up the range. It is usually shown as two lines, %K and its average %D, with 80 and 20 marking overbought and oversold.

What is the stochastic oscillator?

The stochastic oscillator is a momentum indicator popularised by George Lane in the late 1950s. Lane’s observation was simple: in a rising market, prices tend to close near the top of their recent range, and in a falling market near the bottom. When closes start to drift away from the extreme while price is still pushing, the move may be tiring.

So the indicator asks one question of every candle: where did it close compared with the highest high and lowest low of the last 14 candles? If the close equals the highest high, the reading is 100. If it equals the lowest low, the reading is 0. Halfway between gives 50. That is all the raw stochastic is: a position inside a box.

Three things follow from that design, and each one explains how the indicator behaves:

• It is bounded. The reading can never go above 100 or below 0, however far price travels. A pair that rises 300 pips in a week can show the same reading as one that rose 30, as long as each closed near the top of its range.
• It ignores the size of the move. The RSI weighs how big the up-closes were against the down-closes. The stochastic only cares about location. That makes it quicker to reach the extremes.
• It resets when the range resets. As old candles drop out of the 14-candle window, the highest high and lowest low can jump, and the reading can change sharply even on a quiet day.

The name sounds like statistics, but there is nothing random about it. “Stochastic” is just the label that stuck. If you are new to indicators in general, the Academy lesson on indicators explains how they are all built from past prices and why every one of them lags. The new Technical Indicators in Depth course covers this one in its lesson on the stochastic oscillator, and starts with leading and lagging indicators, which is where the stochastic’s reputation as a “leading” tool gets tested.

The %K and %D formula, worked by hand

The raw stochastic line is called %K:

%K = (close − lowest low of n candles) ÷ (highest high of n candles − lowest low of n candles) × 100

The usual look-back, n, is 14. The second line, %D, is a 3-period simple average of %K. It is slower and acts as a signal line, the same idea as the signal line on MACD.

Here are five illustrative daily EUR/USD readings. The highest high and lowest low are for the 14 days ending on each row:

Day	14-day high	14-day low	Close	Close − low (pips)	Range (pips)	%K
1	1.0950	1.0750	1.0900	150	200	75.0
2	1.0965	1.0750	1.0940	190	215	88.4
3	1.0965	1.0750	1.0880	130	215	60.5
4	1.0965	1.0762	1.0915	153	203	75.4
5	1.0965	1.0762	1.0850	88	203	43.3

Check the arithmetic on day 2. The new high at 1.0965 widens the range to 215 pips, and the close at 1.0940 is 190 pips above the low: 190 ÷ 215 × 100 = 88.4. On day 4, the candle that set the 1.0750 low has dropped out of the window, so the lowest low rises to 1.0762. The range shrinks to 203 pips, which is why a close of 1.0915 scores 75.4 rather than something lower.

Now the %D line. On day 3 it is the average of the last three %K values: (75.0 + 88.4 + 60.5) ÷ 3 = 223.9 ÷ 3 = 74.6. On day 4: (88.4 + 60.5 + 75.4) ÷ 3 = 74.8. On day 5: (60.5 + 75.4 + 43.3) ÷ 3 = 179.2 ÷ 3 = 59.7.

Notice how differently the two lines move. %K swung from 88 to 60 to 75 to 43 in four days. %D drifted from 74.6 to 74.8 to 59.7. That smoothing is the reason traders watch the two together: %K reacts first, and %D confirms that the reaction lasted more than one candle.

Fast, slow and full stochastic settings

The version above, raw %K plus its 3-period average, is the fast stochastic. Lane’s raw line jumps around so much that most traders use a smoothed version instead, and this is where the naming gets confusing.

• Fast stochastic: fast %K is the raw calculation; fast %D is its 3-period average.
• Slow stochastic: slow %K is the 3-period average of raw %K (so it equals fast %D); slow %D is a 3-period average of slow %K.
• Full stochastic: the same structure, with all three numbers adjustable: the look-back, the smoothing applied to %K, and the averaging for %D.

Using the table again, the slow %K values on days 3, 4 and 5 are the fast %D values we already worked out: 74.6, 74.8 and 59.7. Slow %D on day 5 is their average: (74.6 + 74.8 + 59.7) ÷ 3 = 209.1 ÷ 3 = 69.7. Two layers of averaging turn a line that jumped 45 points in four days into one that moved about 5.

Why 14, 3, 3 is the default, and why yours may differ

14, 3, 3 is the setting you will see in most textbooks: a 14-candle look-back, 3-period smoothing of %K and a 3-period %D. It is a full stochastic that behaves like the classic slow one. Platform defaults are not consistent, though. Some open the indicator at 5, 3, 3, which is far more sensitive; others at 14, 1, 3, which is the fast version with no smoothing at all. Before you compare your chart with anyone else’s, open the settings and check the three numbers.

A shorter look-back, such as 5, makes the indicator hit 80 and 20 constantly. A longer one, such as 21, makes it calmer but slower to turn. As with any indicator setting, the risk is fiddling: testing combinations until one fits the last few months. Pick a standard setting, write it in your plan, and judge it over many trades.

The timeframe changes the reading more than the setting does. A 14-candle window on a daily chart covers nearly three weeks; on a 15-minute chart it covers three and a half hours. The same pair can be “overbought” on one and “oversold” on the other at the same moment.

Overbought and oversold: why the stochastic oscillator fails in trends

Lane’s levels were 80 and 20. Above 80 is called overbought, below 20 oversold. The words are the problem. “Overbought” sounds like “too high, about to fall”. All a reading of 85 actually says is that the latest closes were near the top of the last 14 candles’ range.

In a range, that is useful. Price bounces between a floor and a ceiling, and the stochastic swings between the zones in step with it. A reading above 80 near the ceiling is one more piece of evidence that price is at the expensive end of its recent box.

In a strong trend, the same reading means the opposite of what beginners expect. A rising market closes near its highs again and again. That is what an uptrend is. So the stochastic climbs above 80 and stays there, sometimes for weeks, dipping briefly to 70 or 75 on each small pause before returning to the high 90s.

This is the main way the stochastic fails, and it fails often. Every time price closes near the top of a fresh 14-candle range, the reading is high, so a persistent trend produces a persistent overbought reading. Selling each one means selling strength repeatedly. No reliable, widely agreed figure exists for how often an overbought reading is followed by a fall; it depends on the pair, the timeframe and how you define a fall, so be wary of any precise success rate quoted online.

There are two practical fixes:

1. Decide the market type first. Use price structure, a moving average or the lesson on trends to judge whether you are in a range or a trend. Only treat 80 and 20 as turning zones in a range.
2. In a trend, use only the readings that agree with it. In an uptrend, ignore overbought and pay attention to oversold. A dip below 20 during an uptrend is a pullback that may be ending, which is a with-trend idea rather than a reversal call.

Our guide to how to identify a trend shows how to make that first call from swing highs and lows, and the guide to pullback trading covers entries in the direction of a trend.

Stochastic crossover signals

The most common stochastic signal is the crossover: %K crossing its own average, %D.

• A bullish crossover is %K crossing above %D. It carries most weight when both lines are below 20.
• A bearish crossover is %K crossing below %D. It carries most weight when both lines are above 80.

A cross in the middle of the scale, around 50, happens constantly and tells you very little. On a busy chart the two lines may cross several times a day.

A crossover is still only the indicator turning, and the indicator turns whenever the close moves within the box. It does not know where support and resistance are, whether a news release is due, or whether the market has just broken out. Three rules keep crossovers in their place:

• Wait for the candle to close. A crossover drawn mid-candle can vanish by the close, because %K is recalculated with every tick.
• Want a reason on the chart. A bullish crossover below 20 at a tested support level is two pieces of evidence. The same crossover in the middle of nowhere is one weak piece.
• Let price trigger the trade. Some traders wait for the stochastic to rise back above 20, or for a candle to close above the previous candle’s high, before acting. Waiting costs a few pips and filters out some false turns.

Lane himself placed more weight on divergence than on crossovers: price making a new low while the stochastic makes a higher low. It works the same way as RSI and MACD divergence, including its habit of appearing early and repeatedly in strong trends. Our guide to RSI and MACD divergence covers how to draw it and when it fails.

Step through the same trade on the chart:

← PreviousNext →

Illustrative EUR/USD 4-hour range. The stochastic oscillator sets the timing; the range floor sets the stop and the range ceiling the target. In a downtrend, the same crossover would have been a weak reason to buy.

Stochastic vs RSI: what is the difference?

Both are bounded momentum oscillators, both run from 0 to 100, and both use a 14-period default. They measure different things, though, and it shows on the chart.

	Stochastic oscillator	RSI
What it measures	Where the close sits in the high-low range	Average size of up-closes against down-closes
Uses highs and lows?	Yes	No, closes only
Usual levels	80 and 20	70 and 30
Lines	Two (%K and %D)	One
Character	Faster, reaches the extremes often	Smoother, reaches the extremes less often

Because the stochastic only cares about location, a modest rise that closes at the top of the range can score 95, while RSI on the same prices might read 65. Pullbacks push the stochastic to the bottom of its scale quickly; RSI usually dips back towards 50.

In practice, the stochastic suits short swings inside ranges, where you want quick readings of “near the top” and “near the bottom”. RSI suits a slower question: is momentum stronger on the up-days or the down-days? Neither predicts the future, and using both does not double your evidence, because both are built from the same candles. Our RSI indicator guide covers the formula and the 50-line trend filter in detail, and our MACD guide covers the moving-average-based alternative.

Switch the playground to RSI and change the period. Short periods behave much like a fast stochastic, hitting the extremes often; longer ones calm down.

SMAEMARSIMACD

Period:

RSI on illustrative prices. Shorten the period and RSI starts to behave like a fast stochastic, touching the extremes often. Lengthen it and the swings shrink. Either way, the indicator is summarising past closes, not forecasting the next one.

What the stochastic oscillator cannot tell you

The stochastic tells you where the close sits in a recent box. It cannot tell you which way the market will go next, and it cannot tell you how far it will go, because the scale is capped at 100. A reading of 98 on the first day of a breakout and a reading of 98 at the end of a long rally look identical.

It also cannot see events. A central-bank decision or a jobs report can carry price straight through a range floor while the indicator shows “oversold” the whole way down. Before you act on any indicator, check the economic calendar for the next few hours; our guide on how to read an economic calendar shows what to look for.

Finally, no indicator changes the basic arithmetic of leveraged trading. When the European Securities and Markets Authority restricted retail CFD leverage in 2018, it cited evidence that between 74% and 89% of retail accounts lost money (ESMA). Leverage magnifies losses as well as gains. Size each trade from its stop, risk a small, fixed share of the account, and use the stochastic as one input among several. The lesson on range trading shows how an oscillator fits into a complete plan with entry, stop and exit.

Common stochastic mistakes

Selling every reading above 80

In a trend, this means selling strength over and over. Overbought describes the recent closes; it is not a forecast. Decide whether the market is ranging or trending before reading the zones.

Acting on a crossover before the candle closes

%K is recalculated with every price change. A crossover that appears at the 40th minute of an hourly candle can be gone by the close. Judge signals on closed candles only.

Comparing readings across different settings

A 5, 3, 3 stochastic and a 14, 3, 3 stochastic on the same chart can disagree by 40 points. Check the settings before comparing your chart with a course, a forum post or a colleague’s screen.

Counting the stochastic and RSI as two confirmations

Both are built from the same recent prices. When they agree, you have one opinion stated twice. Look for confirmation that comes from somewhere else: a level, a candle pattern, a higher-timeframe trend.

Letting the indicator set the stop

The stochastic says nothing about where a trade idea is wrong. Put the stop beyond the level that would prove the idea wrong, then size the trade so that a hit costs your planned amount.

Frequently asked questions

What is the stochastic oscillator in simple terms?

It is a 0 to 100 score of where the latest close sits within the highest high and lowest low of the last 14 candles. A reading of 100 means price closed at the top of that range, 0 at the bottom. It is drawn as two lines, %K and its slower average %D.

What are the best stochastic settings?

No setting is best for every market. The common textbook setting is 14, 3, 3, which gives a smoothed slow stochastic. Shorter look-backs, such as 5, react faster but give many more false signals. Pick one standard setting, check your platform's default, and test it over many trades before relying on it.

What does it mean when the stochastic is above 80?

It means recent closes have been near the top of the last 14 candles' range. In a sideways market that can warn that price is near a ceiling. In a strong uptrend the stochastic can stay above 80 for weeks while price keeps rising, so a high reading on its own is not a reason to sell.

Is the stochastic better than RSI?

Neither is better; they measure different things. The stochastic tracks the close's position in the high-low range and swings to the extremes quickly. RSI compares the size of up-closes and down-closes and moves more smoothly. Many traders find the stochastic more useful in ranges and RSI more useful for judging trend strength.

What is the difference between fast and slow stochastic?

The fast stochastic plots the raw %K line and its 3-period average. The slow stochastic smooths raw %K with a 3-period average first and then averages that again for %D. The slow version gives fewer, later signals. A full stochastic lets you set all three numbers yourself.