An oscillator is a technical indicator that swings back and forth, either between fixed limits or around a center line, to show momentum: how fast and how strongly price is moving. The three most widely used are the RSI, the Stochastic oscillator and the MACD, and each measures momentum in a slightly different way.
Oscillators sit in a panel below the price chart and are calculated only from past prices, so they cannot know anything price does not. Their value is in making momentum easier to see, especially when a move is starting to tire.
What an oscillator measures
Momentum is the speed of price change. A strong trend covers a lot of ground quickly; a tiring trend still moves, but each push covers less. Oscillators turn that into a single line. There are two kinds:
- Bounded oscillators move between fixed limits, usually 0 and 100. RSI and Stochastic are bounded, so they have standard overbought and oversold zones near the extremes.
- Unbounded oscillators have no fixed limits and swing around a zero line. MACD is unbounded, and its values are in price units, so a reading on gold cannot be compared with one on EUR/USD.
Overbought means price has risen far and fast relative to its recent history; oversold means the opposite. Neither means price has to turn.
RSI: the Relative Strength Index
Wilder's RSI compares the size of recent gains with the size of recent losses, by default over 14 candles, and scales the result from 0 to 100.
RSI = 100 − 100 ÷ (1 + RS), whereRS = average gain ÷ average loss
For example, suppose that over the last 14 candles on BTC/USD the average gain works out to $600 and the average loss to $400. RS = $600 ÷ $400 = 1.5, so RSI = 100 − 100 ÷ 2.5 = 100 − 40 = 60. Gains have outweighed losses, but not to an extreme.
Standard readings: above 70 is overbought, below 30 is oversold, and 50 is the midpoint where average gains equal average losses. The RSI guide covers the full calculation, how RSI ranges shift in trends, and hidden divergence.
The Stochastic oscillator
The Stochastic shows where the latest close sits within the high-low range of recent candles. The idea is that in an uptrend closes tend to cluster near the top of the range, and in a downtrend near the bottom.
%K = (close − lowest low) ÷ (highest high − lowest low) × 100
The default setting is 14,3,3:
- 14: the look-back period for the highest high and lowest low.
- 3: the raw value is smoothed with a 3-period average to give %K (the "slow" Stochastic).
- 3: %D is a 3-period average of %K and acts as a signal line.
Above 80 is overbought and below 20 is oversold. Traders watch %K crossing %D, especially inside those zones.
For example, if EUR/USD's highest high over the last 14 candles is 1.0900, the lowest low 1.0800 and the latest close 1.0880, the raw value is (1.0880 − 1.0800) ÷ (1.0900 − 1.0800) × 100 = 0.0080 ÷ 0.0100 × 100 = 80. The close sits 80% of the way up the range, right at the overbought line.
With default settings the Stochastic usually reacts faster than RSI, which means more signals and more false ones.
MACD: Moving Average Convergence Divergence
MACD tracks the gap between two exponential moving averages, or EMAs (see moving averages). The default is 12,26,9:
- MACD line = 12-period EMA − 26-period EMA.
- Signal line = 9-period EMA of the MACD line.
- Histogram = MACD line − signal line.
When MACD is above zero, the 12 EMA is above the 26 EMA and short-term momentum points up; below zero, it points down. A shrinking histogram shows momentum slowing, even while price still edges in the same direction.
For example, if gold's 12 EMA is $2,412 and its 26 EMA is $2,400, MACD = $2,412 − $2,400 = +12. With the signal line at +9, the histogram is 12 − 9 = +3: upward momentum that is stronger than its recent average. If the histogram slips to +1 over the next few candles while price grinds higher, momentum is fading.
Some platforms, including the built-in MACD in MetaTrader 5 (MT5), draw the MACD line itself as bars and the signal line as a line, with no separate histogram. Check which version you are looking at.
Overbought, oversold and divergence
Overbought is not a sell signal
In a strong trend, RSI can stay above 70, or the Stochastic above 80, for a long time while price keeps climbing. Treating every overbought reading as a reason to sell in an uptrend means repeatedly trading against the trend. In a trend, overbought readings usually confirm strength. They carry more meaning in a range, near resistance, or when momentum stops confirming new highs. The same logic applies in reverse to oversold readings in a downtrend.
Regular divergence
Divergence is a disagreement between price and the oscillator:
- Bearish divergence: price makes a higher high, the oscillator a lower high. The new high came with less momentum.
- Bullish divergence: price makes a lower low, the oscillator a higher low. Selling pressure is weakening.
For example, EUR/USD rallies to 1.0950 with RSI at 74, pulls back, then makes a higher high at 1.0980 while RSI peaks at only 66. Price is 30 pips higher (1.0980 − 1.0950 = 0.0030), but momentum is weaker. Divergence is a warning, not a timing tool: it can repeat over several swings while the trend continues, so many traders wait for price to confirm, such as a close below the last swing low.
Combining oscillators with trend and support and resistance
Oscillators work best as one piece of evidence among several. A common sequence:
- Trend: read the direction on a higher timeframe with moving averages or market structure.
- Location: wait for price to reach a meaningful level, such as support and resistance.
- Momentum: check whether the oscillator shows the pullback losing steam, for example RSI turning up from the 30 to 40 area in an uptrend.
- Trigger: wait for price itself to confirm, such as a candle closing back above the level.
An oscillator reading that lines up with the trend and a key level carries far more weight than one in the middle of nowhere.
RSI vs Stochastic vs MACD compared
| Feature | RSI | Stochastic | MACD |
|---|---|---|---|
| Default settings | 14 | 14,3,3 | 12,26,9 |
| Range | 0 to 100 | 0 to 100 | Unbounded, around zero |
| Key levels | 70 and 30, plus 50 | 80 and 20 | Zero line |
| Measures | Average gains vs average losses | Close within the recent high-low range | Gap between two EMAs |
| Speed | Medium | Fast | Slower |
| Common uses | Overbought and oversold, divergence, 50 as a trend filter | Turns within ranges, pullback timing | Trend momentum, crossovers, divergence |
Common mistakes
- Stacking several oscillators. RSI, Stochastic and MACD are all built from the same prices, so three agreeing is not three confirmations. Fix: use one oscillator alongside trend and level.
- Using defaults blindly. Defaults are starting points; 14,3,3 behaves very differently on a 1-minute chart than on a daily chart. Fix: test a setting on the market and timeframe you trade, then keep it fixed.
- Taking signals against the trend. Fix: in an uptrend, focus on oversold or mid-level readings during pullbacks rather than overbought readings at the highs.
- Acting on divergence alone. Fix: wait for price structure, such as a broken swing low or trendline, to confirm.
Key takeaways
- An oscillator measures momentum, either between fixed limits (RSI, Stochastic) or around zero (MACD).
- Defaults: RSI 14 with 70/30, Stochastic 14,3,3 with 80/20, MACD 12,26,9.
- Overbought and oversold describe momentum, not turning points, and can persist in strong trends.
- Regular divergence warns that momentum is fading but needs price confirmation.
- Oscillators work best combined with trend direction and key levels.
Frequently asked questions
Which oscillator is best for beginners?
RSI is usually the easiest place to start: one line, a fixed 0 to 100 scale and widely known levels. Learn one oscillator well on your own market before adding another.
Is MACD an oscillator?
Yes. MACD is an unbounded oscillator: it moves around a zero line rather than between fixed limits. Because it is built from moving averages, it is also described as a trend-following momentum indicator.
Should I use RSI and Stochastic together?
They measure similar things from the same prices, so using both usually adds clutter rather than new information. Many traders pick one momentum oscillator and pair it with a trend tool, such as a moving average, instead.
What timeframe works best for oscillators?
Oscillators work on any timeframe, but lower timeframes produce more readings and more noise. Whatever timeframe you use, read the oscillator in the context of the higher-timeframe trend.