RSI Reversal

Yes, in strong trends RSI can stay in extreme territory for extended periods. This is called 'RSI failure' in mean reversion terms. In a strong uptrend, RSI can stay above 70 for weeks. This is why trend context matters - don't automatically short just because RSI is above 70 in a strong uptrend.

It depends on your approach. Conservative: Wait for RSI to cross back through the threshold (cross above 30 or below 70). This confirms reversal is starting. Aggressive: Enter at the touch, using a tight stop. This gets better entry but more false signals. Moderate: Require a confirming candlestick pattern at the extreme.

Both are momentum oscillators (0-100). RSI compares average gains to average losses over a period. Stochastic compares closing price to the recent price range. Stochastic is typically more sensitive and includes a signal line (%D) for crossover signals. RSI is simpler and often smoother. Many traders use both together for confirmation.

The standard is 14 periods, developed by Wilder. Use this as your baseline. For faster signals (day trading), try 7-9 periods. For smoother signals (position trading or noisy instruments), try 21-25 periods. Test on your specific instrument before changing from the default.

Generally yes, especially for daily charts. RSI is calculated on closing prices. An RSI reading mid-bar might change by close. Waiting for the close confirms the signal. For intraday trading with shorter timeframes, some traders enter before close but use tighter stops.

For bullish divergence: (1) Mark two price lows where the second is lower. (2) Check RSI at both lows - the second RSI reading should be higher. (3) Draw a line connecting the RSI lows - it should slope upward while price slopes down. For bearish divergence, reverse the process. Use clear, distinct swings - minor fluctuations don't count.

In ranging markets, price oscillates between support and resistance without trending. Overbought/oversold conditions reliably lead to reversions to the mean. In trending markets, the trend can override RSI signals - price keeps moving in the trend direction despite extreme RSI. Always check ADX; if ADX > 25, be cautious with RSI reversals.

Multiple touches can mean: (1) Strong trend - RSI keeps getting pushed back to extreme. Be cautious. (2) Building divergence - if price makes lower lows but RSI makes higher lows, this is bullish. Wait for the divergence to complete and RSI to break out of the extreme zone. Don't keep adding to losing positions.

It depends on conditions. 30/70 (standard): More signals, works well in ranging markets. 20/80 (extreme): Fewer signals, but more reliable when they occur, better for trending markets or volatile instruments. You can also use 25/75 as a middle ground. Backtest both on your instrument to see which performs better.

Common approaches: (1) RSI target: Exit when RSI reaches 50 (centerline) or opposite extreme (30→70). (2) Price target: Recent swing high/low or key support/resistance. (3) Partial exits: Take half at RSI 50, trail rest to opposite extreme. (4) Time-based: If RSI doesn't reach target in X bars, exit. RSI 50 is the most common first target.

Adaptive RSI adjusts parameters based on conditions. For period: Calculate ratio of average ATR to current ATR, multiply by base period. High volatility → longer period. For thresholds: Use ADX; high ADX → wider thresholds (80/20). Code this as custom indicator. Backtest with walk-forward validation to ensure adaptation adds value over static parameters.

Classification works well: predict whether the RSI signal will be profitable (1) or not (0). Use ensemble methods (Random Forest, XGBoost) with features: RSI level, RSI slope, divergence flag, ADX, volume ratio, higher TF RSI, distance to S/R. Train on historical signals, validate with walk-forward testing. Set probability threshold (e.g., >60%) to filter signals.

Professionals typically use RSI as one component in multi-factor models, not in isolation. They combine RSI with trend, volatility, volume, and cross-asset signals. They rigorously backtest across instruments and regimes. They often use adaptive parameters and regime-switching logic. Position sizing is based on signal strength and correlation. Everything is validated out-of-sample.

For portfolio-level RSI: (1) Calculate RSI for each asset - overweight assets with bullish RSI context (above 50 but not extreme), underweight extreme or bearish. (2) Use broad market RSI (SPY) as regime indicator - reduce risk when SPY RSI extreme. (3) Combine individual RSI signals with portfolio heat management - don't take too many correlated reversal trades simultaneously.

Key practices: (1) Use standard parameters (14, 30/70) as baseline - only optimize if there's a clear reason. (2) Test that nearby parameters perform similarly (robustness check). (3) Use walk-forward optimization: optimize on period 1, test on period 2, repeat. (4) Keep rules simple - complex rules overfit. (5) If only one specific setting works, it's likely noise. Accept somewhat lower backtested performance for real-world robustness.