Momentum indicators

Williams %R Indicator: Formula, Overbought and Oversold Levels, and a Real S&P 500 Example

Larry Williams’ %R places the close within the high-low range of the last 14 bars on a 0 to −100 scale; here is the formula, how to read −20 and −80, and what those levels did on the S&P 500 in 2025 and since 1970.

· Founder of PatternsWizard
Checked 6 min read
Williams %R(14) on S&P 500 ETF, dividend-adjusted prices, daily, March 20, 2025 to June 13, 2025. Illustration of how the indicator reads, not a backtest.
On this page
  1. How Williams %R is calculated
  2. Overbought, oversold and how to read them
  3. Williams %R on a real chart
  4. Best Williams %R settings
  5. Does it work?
  6. Frequently asked questions

Williams %R is a momentum oscillator that shows where the close sits within the high-low range of the last 14 bars, on a scale from 0 to −100. A reading of 0 means the bar closed at the highest high of the period, −100 at the lowest low. Developed by Larry Williams, it is, as StockCharts puts it, the inverse of the Fast Stochastic Oscillator: the same measure read from the top of the range instead of the bottom.

Williams %R(14) on the daily S&P 500 ETF

Williams %R(14) on S&P 500 ETF, dividend-adjusted prices, daily, December 19, 2024 to June 13, 2025. Illustration of how the indicator reads, not a backtest.
  • Price, daily candles (hollow up, filled down)
  • Williams %R(14) (solid line)
Williams %R(14) on S&P 500 ETF, dividend-adjusted prices, daily, December 19, 2024 to June 13, 2025. Illustration of how the indicator reads, not a backtest.

How Williams %R is calculated

Take the highest high and the lowest low of the last 14 bars, subtract the close from the highest high, divide by the range and multiply by −100. When the range is zero, the value is undefined and our chart leaves the bar blank.

Williams %R: where the close sits in the range

Williams %R: where the close sits in the range: fourteen candles between dashed lines at their highest high and lowest low. A scale on the right is aligned with that range: 0 at the highest high, minus 100 at the lowest low, with the overbought zone from 0 to minus 20 and the oversold zone from minus 80 to minus 100. The last close sits in the top fifth of the range and reads minus 20.
Schematic, not market data.The %R scale is the 14-bar range turned upside down: 0 at the highest high, −100 at the lowest low.

Seen this way, %R is only a position: the close’s place between the highest high and the lowest low, measured down from the top. It does not look at how the price got there, so a close at the top of a range reads the same whether the range was built by a steady rise or a single spike.

Overbought, oversold and how to read them

The traditional thresholds are −20 and −80. Readings from 0 to −20 are called overbought and mean the close is near the top of its 14-bar range; readings from −80 to −100 are called oversold and mean it is near the bottom. StockCharts is explicit that overbought readings are not necessarily bearish and oversold readings are not necessarily bullish: in an uptrend, the ability to keep reaching −20 is a show of strength. It also reads a move back through −50 as a change in short-term momentum.

ReadingWhat it says
0 to −20Close in the top fifth of the 14-bar range
Around −50Close in the middle of the range
−80 to −100Close in the bottom fifth of the range
Repeated readings above −20Closes keep landing near the highs: strength, not an automatic sell

Williams %R on a real chart

The chart under the introduction shows Williams %R(14) on the daily S&P 500 ETF from December 19, 2024 to June 13, 2025, with dividend-adjusted prices. The window opens on an oversold reading, −98.9 at a close of 572.88. Three moments of the window show what the two thresholds do and do not say.

First, oversold was not a floor. On March 14, 2025, %R climbed back above −80 at a close of 551.97, leaving oversold territory. Over the next seven sessions the ETF climbed to 566.07 on March 25, 2.6% higher, while %R reached −9.8, an overbought reading. It then fell, to a close of 488.38 on April 8, 11.5% below the exit from oversold.

Second, the thresholds depend on the range, not on the trend: on April 7, the close of 496.15 was lower than the day before, yet %R rose to −76.1 and left oversold territory because that day’s intraday low had stretched the range.

Williams %R: a lower close that reads higher

Williams %R: a lower close that reads higher: candles fall from the highest high of the 14-bar range. On day 1 the close sits near the lowest low, so Williams %R is near minus 100. On day 2 the close is lower still, but a long lower wick sets a new, much lower low; the range is now taller and the close sits above its minus 80 level, so %R rises out of oversold territory.
Schematic, not market data. Drawn by hand.The case described above: a lower close can lift %R when the low of the day stretches the range.

Third, overbought was not a top. From April 23 to May 20, %R stayed at or above −20 for 20 trading days in a row while the ETF rose from 526.68 to 583.18, a gain of 10.7%; the first close back below −20, on May 21 at 573.35, was still 8.9% above the start of the streak. Over the 120 days of the window, %R was overbought on 35.8% of the days and oversold on 13.3%.

On daily Bitcoin, from June 1 to October 28, 2026, the split was more even: %R(14) was overbought on 20.7% of the 150 days and oversold on 16.7%. The longest overbought run lasted 11 days, from August 19 to August 29, while Bitcoin rose 12.8%. Here too, the reading that traders call overbought marked the strongest stretch of the window, not its top. Reading overbought as a sell signal is a bet on mean reversion: some markets and periods revert, others trend.

Williams %R(14) on daily Bitcoin

Williams %R(14) on Bitcoin, daily, June 1, 2026 to October 28, 2026. Illustration of how the indicator reads, not a backtest.
  • Price, daily candles (hollow up, filled down)
  • Williams %R(14) (solid line)
Williams %R(14) on Bitcoin, daily, June 1, 2026 to October 28, 2026. Illustration of how the indicator reads, not a backtest.

Best Williams %R settings

14 bars is the default. A shorter setting makes the oscillator more sensitive and a longer one less so, as StockCharts notes. With 10 bars on the S&P 500 index since 1970, %R was overbought on 33.5% of the days and oversold on 16.2%, close to the 14-bar figures of 35.1% and 15.3%. On this index, both settings spent about twice as long overbought as oversold.

Share of days overbought and oversold, S&P 500 index

Share of days overbought and oversold, S&P 500 index. Overbought (−20 or above): 10 bars 33.5%, 14 bars 35.1%; Oversold (−80 or below): 10 bars 16.2%, 14 bars 15.3%. Our computation; the inputs and the method are in the text.
  • 10 bars
  • 14 bars
Our computation; the inputs and the method are in the text.Daily bars from January 1970 to October 28, 2026.

Does it work?

Our bench tests chart patterns detected by rule against random entries, not indicators (see our methodology). We measured the two threshold crossings of Williams %R(14) on the daily S&P 500 index from January 1970 to October 2026, without dividends or costs. Over the 20 trading days after a move back above −80, the index rose 62.3% of the time, with a mean change of +0.83%, across 872 signals. After a move back below −20, it still rose 61.2% of the time, with a mean of +0.65%, across 1,155 signals. All days together gave 61.4% and +0.73%. Neither signal separated itself from an ordinary day on this index. These are descriptive averages, not a trading test.

Index higher 20 trading days later, S&P 500 index

Index higher 20 trading days later, S&P 500 index. After a move back above −80 (872): 62.3%, mean +0.83%; After a move back below −20 (1,155): 61.2%, mean +0.65%; All days: 61.4%, mean +0.73%. Our computation; the inputs and the method are in the text.
Our computation; the inputs and the method are in the text.Williams %R(14), daily bars from January 1970 to October 28, 2026, without dividends or costs. Number of signals in brackets.

The %R is the mirror image of the stochastic oscillator; compare it with the RSI and the stochastic RSI, or browse all momentum indicators.

Every pattern's results on one sheet

The only pattern cheat sheet with tested stats: 152 patterns, what the textbook says and what really happened on 6 markets. Free PDF, sent to your inbox.

The PDF by email, plus our newsletter: new tests and results that change. Unsubscribe anytime. Privacy.

Frequently asked questions

How do you read Williams %R?

It shows where the close sits in the range of the last 14 bars: near 0 at the top, near −100 at the bottom. Above −20 is called overbought and below −80 oversold, but in a trend the indicator can stay at one extreme for weeks.

Who created Williams %R?

Larry Williams. StockCharts describes it as the inverse of the Fast Stochastic Oscillator.

What is the best setting for Williams %R?

14 bars is the default. Shorter settings are more sensitive, longer ones smoother. No setting has been shown to be the best.

Is Williams %R the same as the stochastic oscillator?

Almost: %R measures the distance from the close to the top of the range, the fast stochastic %K the distance from the bottom. %R equals %K minus 100.

Sources

  1. StockCharts ChartSchool, Williams %R
, Founder of PatternsWizard

Passionate about finance, quantitative trading and data analysis. Amaury built the PatternsWizard test bench, which detects each pattern on full price history and measures what happened next against random entries.

PatternsWizard tests chart and candlestick patterns on full market history, against random entries. How we test patterns.

179 patterns testedRandom-walk self-test passed

Related guides

All momentum indicators articles