Oscillator (Technical Analysis)

An oscillator, in technical analysis, is an indicator calculated from price or volume that moves back and forth around a central level or between fixed upper and lower bounds. Where a price chart can trend indefinitely, an oscillator is constructed so that its values stay within a range, which makes it possible to describe the current reading as high or low relative to the past. Traders use oscillators to judge the strength of a move and to look for signs that it may be running out of force.

Oscillators form one of the two main families of indicators in technical analysis. The other is the trend-following family, such as moving averages, which are plotted on the price chart itself. Oscillators are usually drawn in a separate panel beneath the chart. Most measure some form of momentum: the speed of price change, not its level.

• Banded oscillators. These are scaled to run between fixed limits, usually 0 and 100. The relative strength index and the stochastic oscillator are the best-known examples. Readings near the top of the range are called overbought and readings near the bottom oversold.
• Centred oscillators. These fluctuate above and below a zero line with no fixed limits. Examples are the moving average convergence divergence indicator, the rate of change and the simple momentum indicator, which is the current price less the price a set number of periods ago. A move from negative to positive is read as a shift from falling to rising momentum.

The relative strength index, introduced by J. Welles Wilder in 1978, compares the average size of recent gains with the average size of recent losses:

Here $RS$ is the average gain over the look-back period divided by the average loss over the same period, with 14 periods as the usual setting. If the average gain is 1.5 and the average loss is 0.5, then $RS = 3$ and the index is $100 - 100 / 4 = 75$. Readings above 70 are conventionally treated as overbought and below 30 as oversold.

The stochastic oscillator, popularised by George Lane, shows where the latest close sits within the recent trading range:

• $C$ is the most recent closing price.
• $L_n$ is the lowest low price of the past $n$ periods.
• $H_n$ is the highest high price of the past $n$ periods.

If the close is 48 and the range of the past 14 days runs from 40 to 50, then $\%K = (48 - 40) / (50 - 40) \times 100 = 80$. The thresholds usually applied are 80 and 20.

• Overbought and oversold levels. A reading beyond a threshold suggests that the price has moved far and fast and may pause or reverse. This is the logic of mean reversion applied to short-term price action.
• Centre-line crossovers. A centred oscillator crossing zero, or a banded one crossing its midpoint, is taken as a change in the direction of momentum.
• Signal-line crossovers. Many oscillators are plotted with a smoothed version of themselves, and a cross of the two lines is used as a trading signal.
• Divergence. When price makes a new high but the oscillator makes a lower high, the rise is said to be losing momentum, a warning of a possible reversal. The opposite pattern at lows is a bullish divergence.

How well oscillator signals work depends on whether the market is ranging or trending. In a sideways market, prices swing between a support level and a resistance level, and selling overbought readings and buying oversold ones fits that behaviour. In a strong trend the same rule fails repeatedly: an oscillator can stay overbought for weeks while the price continues to climb, and each sell signal is a loss. Analysts have noted that in an uptrend the relative strength index tends to hold in a higher band.

For this reason oscillators are rarely used alone. A common practice is to determine the trend first, with a moving average or a longer chart interval, and then to take only those oscillator signals that agree with it, for example buying oversold readings within an uptrend. Combining oscillators with Bollinger Bands or with volume is another approach.

Oscillators are transformations of past prices and contain no information that is not already in the price series. Several popular ones are closely related mathematically, so agreement among them is weaker confirmation than it appears. Their thresholds and look-back lengths are conventions, not constants, and different settings give different signals. Divergences are easy to identify afterwards and harder to act on in real time, since a divergence can persist through several further highs.

Evidence on profitability is mixed. Rules based on standard oscillator settings have generally shown little reliable profit in liquid markets once trading costs are included, and results that look strong in a backtest often rely on parameters tuned to that period, a form of overfitting. Oscillators remain widely used as descriptive tools, as inputs to larger systematic models and as filters within discretionary trading, and they are standard features of every charting platform.