RSI Divergence Trading

Visual steps: 1) Look at recent price swings - identify clear highs and lows. 2) Place RSI indicator below price chart. 3) Draw a line connecting two recent price lows (for bullish) or highs (for bearish). 4) Draw a line connecting RSI values at those same points. 5) If lines slope in opposite directions = divergence. For bullish: price line slopes down, RSI line slopes up. For bearish: price line slopes up, RSI line slopes down. Most charting platforms let you draw lines on RSI to make comparison easier.

Yes, divergence works on any timeframe, but characteristics differ: 1-5 minute: many signals, most are noise, low reliability. 15-60 minute: good for intraday, moderate reliability. 4-hour/Daily: best for swing trading, higher reliability. Weekly: major reversals, very reliable but rare. Recommendation: start with hourly or daily timeframe. Shorter timeframes have more false signals. Higher timeframes give fewer but more reliable signals. Match timeframe to your trading style and holding period. SGX two-session trading lets you act on signals across Asian, European and US hours.

No, selective trading is key. Filter divergences by: 1) Trend context - regular divergence best at trend extremes after extended moves. 2) RSI level - should be in overbought/oversold zones. 3) Confirmation - wait for reversal candle. 4) Support/resistance - divergence at key levels is stronger. 5) Higher timeframe alignment - check if bigger picture supports the trade. Trade only high-quality setups. Missing mediocre setups is better than taking losses on weak signals. Quality over quantity.

Key difference is what they signal: Regular divergence: counter-trend signal, suggests reversal. Price makes new extreme but RSI doesn't confirm = trend exhaustion. Hidden divergence: trend continuation signal. Price pulls back but doesn't make new extreme, RSI makes new extreme = trend resuming. Easy distinction: Regular = price at NEW extreme, RSI not confirming. Hidden = price NOT at new extreme (pullback), RSI at new extreme. Regular is riskier (counter-trend), hidden is safer (with-trend).

General guidelines by timeframe: 15-min divergence: should show results in 1-4 hours. Hourly divergence: should show results in 1-2 days. Daily divergence: should show results in 3-10 days. If timeframe-appropriate time passes without progress, reassess: 1) Is the divergence extending (making bigger divergence)? Wait if setup improves. 2) Is it failing (price breaking divergence extreme)? Exit at stop. 3) Is it consolidating? May need patience. Use time stops if price goes nowhere - your capital can be better used elsewhere.

Hidden divergence trading approach: 1) First confirm the trend is intact (higher highs/lows for uptrend). 2) Wait for pullback within the trend. 3) Look for hidden divergence: price higher low + RSI lower low (bullish). 4) Enter when RSI turns back up from its lower low. 5) Stop below the price higher low. 6) Target: new trend highs. Advantage: you're trading with the trend, so higher success rate. Key insight: hidden divergence shows the trend is resuming despite temporary RSI weakness. The market is using pullback to gather energy for continuation.

Dual divergence is powerful: 1) Identify divergence on RSI. 2) Check if MACD shows same divergence (price extreme, MACD not confirming). 3) Both showing same divergence = high probability signal. 4) Trade with higher confidence and potentially larger size. Different information: RSI measures momentum speed. MACD measures trend momentum. Both disagreeing with price = strong exhaustion signal. If they conflict (RSI shows divergence, MACD doesn't): wait for alignment or trade with reduced conviction. Dual confirmation significantly improves win rate.

Top failure causes: 1) Strong trend override - powerful trends can ignore divergence for extended periods. ADX > 40 divergences less reliable. 2) No confirmation - entering just on divergence without waiting for reversal sign. 3) Poor context - divergence in middle of range vs at clear trend extreme. 4) News events - fundamentals can override technical signals. 5) Weak divergence - slight RSI non-confirmation isn't significant. 6) Early entry - entering on first divergence when it extends to larger divergence. Prevention: use filters (trend strength, confirmation, S/R confluence), accept some failure as normal, size appropriately.

Volatility impacts: High volatility: wider stops needed, divergence swings are larger, confirmations are clearer but price moves fast. Low volatility: tighter stops possible, divergence may be subtle, moves are slower to develop. VIX/Nikkei VI context: when the relevant volatility gauge is elevated (e.g. VIX > 25), divergences may be more reliable (extreme sentiment) but use wider stops. When it is low (e.g. VIX < 14), divergences in a complacent market may need more confirmation. Adjustment: in high volatility, wait for stronger confirmation and use ATR-based stops. In low volatility, signals may be weaker but trade more precise levels.

Yes, divergence works with options: Advantages: defined risk (option premium is max loss), potentially higher percentage returns, can use spreads for defined risk/reward. Considerations: theta decay works against you, need sufficient time to expiration, IV impacts premium. Strategy adaptations: buy slightly ITM or ATM options for divergence trades (less theta/IV impact). Use spreads (bull call/put spread) to reduce cost. Allow enough time (minimum 2-3x expected holding period to expiration). Example: bullish divergence on daily Nikkei 225 - buy a 1-month ATM call or a bull call spread (SGX lists options on Nikkei 225 and MSCI Singapore).

System components: 1) Swing detection algorithm (price pivots with minimum X bars between). 2) RSI calculation at each swing. 3) Divergence detection (compare consecutive same-type swings). 4) Magnitude scoring (angle difference, class A/B/C). 5) Filter criteria (ADX threshold, RSI extreme requirement). 6) Entry logic (confirmation candle, RSI turn). 7) Exit rules (target levels, stops, time exits). Backtesting: test on 3+ years data, minimum 100 trades for significance. Walk-forward validate. Key metrics: win rate 45-55% typical, profit factor 1.4-1.8, max drawdown < 20%. Avoid overfitting by keeping rules simple and testing on multiple instruments (e.g. China A50, Nikkei 225, SiMSCI).

Elliott Wave integration: Wave 3: strongest wave, divergence rare (momentum strong). Wave 5: final wave, often shows divergence vs wave 3 (momentum weakening). Divergence in wave 5 = wave completion signal, expect correction. Wave B: corrective wave often shows divergence vs prior trend extreme. Wave C: ending diagonal patterns often have divergence at completion. Practical use: if count suggests wave 5 forming, look harder for divergence. Divergence confirms wave exhaustion. Major divergence on weekly/monthly often marks primary wave endings. Caution: wave counts are subjective, use divergence as confirmation rather than sole signal.

Mathematical limitations: 1) Lookback dependency: RSI value depends on N periods. Different N gives different divergence signals. 2) Smoothing artifact: RSI is bounded 0-100, extreme prices compressed at boundaries. Divergence may reflect compression, not true momentum change. 3) Starting point bias: RSI calculation affected by initial values in lookback window. 4) No statistical edge guarantee: academic studies show mixed results on divergence profitability. 5) Survivorship in examples: memorable divergences that worked, forgetting failures. Mitigation: use multiple confirmation sources, understand divergence as probability enhancement not guarantee, rigorous backtesting before trusting.

ML enhancement approaches: 1) Classification model: features include divergence magnitude, RSI level, volume ratio, trend strength, time factors. Target: did divergence lead to profitable reversal? 2) Feature importance: ML identifies which factors most predict divergence success. 3) Dynamic thresholds: ML adjusts divergence magnitude threshold based on market regime. 4) Pattern recognition: CNN/RNN to identify divergence patterns beyond simple formulas. 5) Ensemble: combine ML prediction with traditional rules. Implementation: start with random forest for interpretability. Use walk-forward validation. Track model decay and retrain periodically. ML should enhance, not replace, fundamental divergence logic. Avoid black box models you don't understand.

Regime performance analysis: Trending regime (ADX > 30): regular divergence often fails or extends. Hidden divergence works well for continuation. Reduce regular divergence trading, focus on hidden. Ranging regime (ADX < 20): regular divergence works well at range extremes. Key reversal indicator in ranges. Increase regular divergence trading. Volatile regime (VIX/Nikkei VI elevated): divergences are clearer (larger swings) but need more confirmation and wider stops. Mixed regime (transition): divergences may signal regime change. Watch for major divergences at regime inflection points. Adaptation: classify current regime first, then select appropriate divergence strategy. Don't use same approach in all regimes.