What are oscillators?

Under many price charts sits a second panel with a line that rises and falls between two limits, often 0 and 100. That line is an oscillator.

It shows where the latest price sits compared with recent prices. The labels traders attach to its extremes often promise more than that.

What does an oscillator measure?

The previous lesson sorted indicators into leading and lagging. Oscillators are often filed in the first group, yet every one of them is calculated from prices that have already printed.

An oscillator is an indicator that turns recent price action into a number moving back and forth within a range or around a centre line.

Most oscillators take a setting called the period, the number of candles in the calculation. A common default is 14.

The stochastic oscillator shows the idea most clearly. Its main line, %K, asks how far up the recent range the last candle closed:

%K = (close minus lowest low) ÷ (highest high minus lowest low) × 100

Here is a worked example with round illustrative prices on EUR/USD, where a pip is 0.0001:

1. Over the last 14 candles, the highest high is 1.10500 and the lowest low is 1.10000.
2. The range is 1.10500 minus 1.10000 = 0.00500, or 50 pips.
3. The last candle closes at 1.10400, which is 0.00400, or 40 pips, above the low.
4. %K = 40 ÷ 50 × 100 = 80.

A reading of 80 means the close sits 80% of the way up the 14-candle range. Had the candle closed at 1.10050 instead, the reading would be 5 ÷ 50 × 100 = 10, close to the bottom.

How do bounded and centred oscillators differ?

Oscillators come in two families. A bounded oscillator cannot leave its range, whatever the price does. The stochastic and the RSI both run from 0 to 100.

A centred oscillator has no ceiling or floor. It swings above and below a middle value, usually zero. MACD is the best known example: it is positive when a faster moving average sits above a slower one.

Williams %R is the stochastic turned upside down. With the same period, it equals the unsmoothed %K minus 100, so the reading of 80 in the example would show as -20.

A centred oscillator has no fixed extreme. Its size depends on how far the price itself moves, so a MACD value that looks large on one pair or timeframe can be ordinary on another.

What do overbought and oversold mean?

On a bounded oscillator, readings near the top are called overbought and readings near the bottom oversold. For the RSI, the usual lines are 70 and 30, the levels its creator J. Welles Wilder suggested.

The stochastic is more often drawn with lines at 80 and 20. On either indicator, the zones describe recent closes and nothing more.

An RSI above 70 says recent closes have been strong over the chosen period. It says nothing about fair value, and nothing about whether buyers have run out.

The lines themselves are conventions. You can move the RSI lines to 80 and 20 so that fewer readings count as extreme. The market does not change when you do this, only the number of signals you see.

Why extremes can last in a strong trend

Go back to the EUR/USD example, where %K read 80 at 1.10400. Suppose the next candle rises and closes at its high of 1.10600, the highest price in the period.

The close is now the top of the range, so close minus lowest low equals the whole range, and %K = 100. If the candle after that closes at its high of 1.10700, %K is 100 again.

A trader who sold at 1.10400 because the reading said overbought would now be 30 pips behind before the spread, since 1.10700 minus 1.10400 = 0.00300. The oscillator would still say overbought.

Picture a steadier uptrend where EUR/USD climbs 100 pips, from 1.10000 to 1.11000. Each dip is shallow, so closes stay near the top of the range and the stochastic stays above 80 for most of the move.

Extreme readings tend to line up with turns more often in a sideways market, where price keeps reversing near the edges of its range. The lesson on trending and ranging markets shows how to tell the two apart.

What traders do with an oscillator once a trend is running is the subject of the next lesson, using oscillators when the market trends.

How the period changes the reading

The period decides how much history the oscillator compares against. A short period uses a narrow range, so each new candle moves the line further and extremes come more often.

Take the last 5 of the same 14 candles. Their highest high is 1.10420 and their lowest low is 1.10260, a range of 16 pips. The close of 1.10400 is 14 pips above that low.

On a 5-candle setting, %K = 14 ÷ 16 × 100 = 87.5, while the same close reads 80 on a 14-candle setting. The price is identical, and only the setting changed.

Most platforms also smooth the stochastic. The common slow version averages %K over 3 candles and adds a second line, %D, which is an average of that smoothed line. Smoothing cuts noise and adds delay.

How traders read oscillators, and how good the evidence is

A handful of readings come up again and again, and each has a known weakness:

• A move back out of a zone, such as RSI falling from above 70 to below it, read as a fading push. In a trend the line often returns to the zone soon after.
• A cross of the middle line, 50 or zero, read as a change in momentum. By the time it happens, much of the move may be over.
• A divergence, where price makes a higher high and the oscillator a lower high. It can repeat over several swings before price turns, or never lead to a turn.
• An extreme reading at a support or resistance level in a range, used as a second opinion. The range can break at any time.

The evidence for simple overbought and oversold rules is weak. Tests of fixed oscillator rules on currencies tend to give results that change with the pair, the settings and the years studied, and trading costs can erase a small edge.

If you want to use an oscillator rule, write it down exactly and backtest it on your own pair and timeframe, costs included. The examples in this lesson illustrate the mechanics and are not recommendations.