Profits from price reverting from overbought/oversold extremes
| Strategy Type | Mean Reversion / Momentum |
| Market Outlook | Profits from price reverting from overbought/oversold extremes |
| Risk Profile | Moderate - Fast oscillator with quick signals |
| Reward Profile | Quick profits from mean reversion bounces |
| Time Horizon | Day trading to swing trading (hours to days) |
| Iv Environment | Works in any IV; often elevated at extremes |
| Breakeven | Entry price +/- stop distance |
| Primary Instruments | SPY, QQQ, DIA (ETFs), Large-cap stocks, Futures, Forex, Crypto |
| Sec Compliance | Standard trading rules; no special requirements |
| Contract Size | 100 shares (stocks), varies by futures contract |
| Trading Hours | 9:30 AM - 4:00 PM ET (stocks), nearly 24 hours (futures/forex/crypto) |
| Expiry Options | N/A - Stock/ETF/Futures strategy (options overlay possible) |
| Settlement | T+1 for stocks/ETFs, same day for futures |
| Margin Requirements | Reg T for stocks (50% initial), varies for futures |
| Pdt Rule | Applicable if day trading with under $25K |
| Tax Treatment | Short-term capital gains for typical holding period |
Larry Williams designed %R to show how far price has fallen from the high of the range. A reading of -20 means price is only 20% below the high (overbought). A reading of -80 means price is 80% below the high (oversold). It's intuitive once you understand: bigger negative number = price further from the high = more oversold.
Very similar but not identical. Both measure where close is relative to the high-low range. The key differences: (1) %R uses inverted scale (-100 to 0), Stochastic uses normal (0 to 100). (2) %R has no smoothing, Stochastic smooths %K and has signal line %D. (3) %R is faster but with more whipsaws.
Generally, wait for %R to turn and cross back above -80 (for longs). Just hitting -80 doesn't guarantee reversal - %R could continue to -90 or -95. The crossback confirms the bounce is starting. You give up a few ticks but reduce the risk of catching a falling knife.
14 is standard and works for most swing trading. For day trading, try 5-10 for faster signals. For position trading, use 20-30 for smoother signals. Always match the period to your timeframe - shorter period for shorter timeframes.
Yes, in strong trends. In a powerful uptrend, %R can stay above -20 for extended periods. In a downtrend, it can stay below -80. This is why trend context matters - don't automatically fade %R extremes in trending markets.
For bullish divergence: Find two price lows where the second is lower. Check %R at both - the second should be higher (less negative, like -85 vs -92). Draw lines: price sloping down, %R sloping up. For bearish: Find two higher price highs with lower %R highs. Divergence is strongest in extreme zones.
%R has no smoothing - it's the raw calculation. Stochastic smooths %K with a moving average and adds signal line %D. This smoothing filters out some noise but adds lag. %R's lack of smoothing means faster signals but more false signals (whipsaws). Choose based on speed vs reliability preference.
Use -10/-90 instead of -20/-80 in: (1) Trending markets where standard extremes give premature signals. (2) Very volatile instruments where -80 is reached frequently. (3) When you want fewer, higher-probability signals. The trade-off is fewer opportunities but higher win rate per signal.
RSI works well since both are momentum oscillators - when both are oversold, it's double confirmation. Stochastic is similar to %R, so their agreement is meaningful. Support/Resistance adds price-level confluence. ADX helps filter - only trade %R reversals when ADX < 25 (ranging market).
Primary target: -50 midline (half the move). Extended target: Opposite extreme (-20 for longs from -80, -80 for shorts from -20). Many traders take partial profits at -50 and trail the rest. You can also use price-based targets (support/resistance) coordinated with %R levels.
Adjust period based on volatility: Period = 14 × (Current ATR / Average ATR). Adjust thresholds based on trend: If ADX > 25, use -10/-90; if ADX < 20, use -20/-80. Code as custom indicator with dynamic parameters. Backtest against static system to validate improvement. Use walk-forward testing.
Classification to predict whether signal reaches target works well. Features: %R value, slope, time in zone, divergence flag, ADX, RSI, volume ratio, higher TF %R. Use Random Forest or XGBoost. Set probability threshold (>60%) to filter. Walk-forward validate. Retrain monthly as market changes.
Professionals use %R as one factor in multi-factor models, not in isolation. They combine with trend, volatility, and fundamental factors. Everything is backtested across instruments and regimes with walk-forward validation. Position sizing may scale with %R extremity. They monitor for alpha decay and adapt.
Key limitations: (1) No signal line - relies purely on threshold crosses. (2) Very volatile/whippy - many false signals. (3) Can stay extreme in trends. (4) Identical to inverted Stochastic %K - no unique information. (5) Fixed thresholds may not suit all instruments. Use with filters and confirmation.
At portfolio level: Track %R for broad market (SPY) as regime indicator. When SPY %R is extreme, market may be due for reversal. For individual positions: Aggregate %R signals - if many holdings are simultaneously overbought, consider reducing exposure. Use %R as timing tool for portfolio adjustments.
Full guided lessons, quizzes, and a complete strategy library for the United States market. One-time purchase. No subscription, ever.
Get United States access →