Williams %R

Williams %R is the trading-world equivalent of left-handed math. The formula and the indicator do exactly what Stochastic does, but with the signs flipped and the scale rearranged. It exists primarily because Larry Williams, the futures trader who developed it in the early 1970s, preferred reading the indicator that way. The substance is identical to Stochastic; the presentation is different.

Whether you use Williams %R or Stochastic is essentially a preference for which direction the scale runs. The signals they produce are mathematically equivalent. Running both is redundant.

What Williams %R actually measures

Williams %R answers the same question as Stochastic: where does the current close sit within the recent high-low range?

The difference is purely cosmetic. Stochastic places "high in the range" at 100 (a positive number near the top of the scale). Williams %R places "high in the range" at 0 (the highest reading on a -100-to-0 scale). The math is identical; the visual presentation is inverted.

If the most recent close is at the top of the recent range, Williams %R reads near 0. If at the bottom, it reads near -100. If in the middle, -50.

The "overbought" zone in Williams %R is above -20 (closer to 0); the "oversold" zone is below -80 (closer to -100). The thresholds are the inverse of Stochastic's 80 and 20.

The formula

%R = ((Highest High - Close) / (Highest High - Lowest Low)) × -100

where Highest High = highest high over the last N periods
      Lowest Low = lowest low over the last N periods

Compare this to Stochastic's %K:

%K = ((Close - Lowest Low) / (Highest High - Lowest Low)) × 100

The numerator is reversed (Highest High - Close instead of Close - Lowest Low) and the scaling constant is negative. The two indicators sum to 100 by construction: %K + |%R| ≡ 100.

Default lookback is 14 periods, matching Stochastic's default %K period.

A worked example

Take the same scenario as the Stochastic example. Over the last 5 bars, highest high is 110, lowest low is 90, current close is 100 (mid-range).

%R = ((110 - 100) / (110 - 90)) × -100
   = (10 / 20) × -100
   = 0.5 × -100
   = -50

Williams %R reads -50, mid-range. (Stochastic on the same data reads 50 - the inverse on its 0-100 scale.)

If price rallies to 108:

%R = ((110 - 108) / (110 - 90)) × -100
   = (2 / 20) × -100
   = -10

%R moves toward 0 (above -20 = overbought zone). On Stochastic, %K would have moved to 90 (also overbought).

If price drops to 92:

%R = ((110 - 92) / (110 - 90)) × -100
   = (18 / 20) × -100
   = -90

%R moves toward -100 (below -80 = oversold zone). On Stochastic, %K would be at 10.

The two indicators describe the same data with the scales flipped.

How traders actually use Williams %R

The setups are identical to Stochastic. The label changes; the trade does not.

1. Range-bound mean reversion

In a sideways market between defined support and resistance:

• %R reaches above -20 (overbought) → short setup forming at top of range
• %R crosses back below -20 → entry trigger
• Stop above range high; target range low

Mirror image for longs at -80. Works only in clean ranges; fails in trends.

2. Divergence

Bullish divergence: price prints a lower low, %R prints a higher low (closer to 0). Bearish divergence: price prints a higher high, %R prints a lower high (further from 0). Same logic as RSI / Stochastic divergence.

3. Midpoint cross as trend filter

%R crossing above -50 from below indicates closes have shifted to the upper half of the recent range - bullish posture. Crossing below -50 indicates bearish posture.

This is the cleanest %R use. Treat the -50 line as a trend filter; ignore the -20 and -80 zones in favor of divergence reads.

The trap most retail traders fall into

Same trap as Stochastic and RSI: treating -20 and -80 as automatic sell and buy signals in trending markets.

In a strong uptrend, %R pins above -20 for extended stretches. Each cross below -20 (only to recross back) gets interpreted as a sell signal. The result is a series of losing short trades against the dominant trend.

The fix is the same: regime filter first. Only act on %R extreme readings in genuinely range-bound markets. In trends, use divergence and the -50 cross instead.

The second %R-specific trap: confusion from the inverted scale. Traders sometimes misread -20 as "very negative, must be oversold" because of the minus sign. The opposite is true: -20 is near the top of the scale (close to 0), which is overbought. Get familiar with the inversion before trading off it.

The third trap: using both %R and Stochastic on the same chart. They are mathematically equivalent. Plotting both does not provide additional information; it provides the same information twice.

Williams %R vs other momentum indicators

vs Stochastic. Mathematically equivalent. Pick one. Williams %R has slightly less noise because there is no signal-line smoothing layer (no %D equivalent); Stochastic offers the smoothed %D for crossover signals. If you want the smoothed crossover, use Stochastic; if you prefer the raw oscillator, use Williams %R.

vs RSI. Both are bounded oscillators with similar intent (overbought / oversold). RSI is built from the gains/losses ratio; %R is built from range position. They are correlated but not identical. RSI handles trends slightly better; %R is faster.

vs CCI. %R is bounded (-100 to 0). CCI is unbounded. %R is preferred for clean range-bound setups; CCI for capturing the magnitude of extreme moves.

Common questions

Why is Williams %R negative? Because Larry Williams chose to present it that way - subtracting the close from the recent high (rather than subtracting the recent low from the close) and applying a negative scaling factor. The math is mathematically equivalent to Stochastic; the presentation is preference.

Should I use the 14-period default? Yes. 14 matches Stochastic, RSI, and most of Wilder's indicators. The convention exists for a reason.

Why does Williams %R look "upside down" compared to Stochastic? Because the scale runs -100 (bottom) to 0 (top), where Stochastic runs 0 (bottom) to 100 (top). The two oscillators are mirror images on their respective scales.

Is Williams %R better than Stochastic? No - they are mathematically equivalent. The choice is preference. Williams %R has slightly less smoothing (no %D equivalent), which some traders prefer. Most retail platforms feature both; few traders use both simultaneously.

Does Williams %R work on intraday? Yes, with the same caveats as Stochastic: 15-minute and higher timeframes produce more reliable signals; 1- and 5-minute charts have too much noise.

Can Williams %R predict reversals? Only via divergence, and only with confirmation from structure. Extreme readings by themselves are not predictive in trending markets.

When to use Williams %R and when not to

Use Williams %R when:

• You prefer the inverted scale to Stochastic's standard scale
• You want a bounded oscillator without an internal smoothing layer
• You are looking for range-bound mean-reversion or divergence setups

Skip Williams %R when:

• You are already using Stochastic - the two are equivalent
• You are in a strong trend and tempted to fade extreme readings
• You are unfamiliar with the inverted scale - it produces confusion when read alongside positive-scale indicators