Aluminium RSI Strategy

RSI works well for Aluminium because the metal has moderate volatility that leads to meaningful mean reversion after extreme readings. Unlike highly volatile metals like Nickel where RSI can stay extreme for extended periods, Aluminium's steadier price action means oversold and overbought readings more reliably lead to reversals. The moderate daily moves (0.5-1.5%) allow RSI signals to play out without excessive noise.

RSI works alone but combining with filters improves results significantly. The most effective addition is a 50 EMA trend filter - only take buy signals above 50 EMA, sell signals below. This simple filter can improve win rate from ~50% to ~58-62%. Avoid adding too many indicators; one momentum indicator (RSI) plus one trend filter (EMA) is usually sufficient. More indicators often lead to analysis paralysis.

For intraday trading, 15-minute charts work well, providing enough signals without excessive noise. For positional trades (multiple days), use daily charts. Always check higher timeframe RSI for context - daily RSI above 50 means bullish bias for intraday trades. The 30-minute timeframe is a good middle ground for swing traders holding 1-3 days.

RSI signals are more reliable when: (1) They occur at key support/resistance levels (confluence), (2) Higher timeframe RSI supports the trade direction, (3) Volume/liquidity confirms the move, (4) SHFE Aluminium and the broader base-metals complex are moving in the same direction. Avoid RSI signals during low-liquidity periods (very early Asian hours before European liquidity builds), random price levels, or against the major trend.

Your stop loss should protect you. RSI can stay overbought/oversold during strong trends - this is why stops are essential. If stopped out, review the trade: was there confluence? Was the higher timeframe supportive? Was it during a valid trading session? Learn from each loss to improve future signal selection rather than abandoning the strategy.

For effective divergence trading: (1) Only trade divergences at significant support/resistance levels, (2) Wait for price confirmation - a strong reversal candle after divergence forms, (3) Trade divergence in direction of higher timeframe trend for best results, (4) Set stop beyond the extreme that created the divergence. Remember: divergence shows weakening momentum but doesn't guarantee reversal - confirmation is essential.

In high volatility (ATR elevated): consider shorter RSI period (10-12) for faster signals or raise overbought/oversold levels to 75/25 to filter noise. In low volatility: standard 14-period works well, can even extend to 18-21 for cleaner signals. In trending markets: watch for range shift and adjust your OB/OS expectations (40-80 in uptrends, 20-60 in downtrends).

Very important. The LME is the global benchmark, but SHFE (Shanghai) is the other major hub and the LME-SHFE arbitrage drives physical and speculative flows. If LME RSI is oversold but SHFE is trending down strongly, the bounce may be limited. Best signals occur when both LME and SHFE RSI are oversold and showing signs of reversal. During the Asian session, monitor SHFE closely; during European and US hours, weight the Dollar Index and COMEX. Signals without any intermarket confirmation should be treated with more caution.

Range shift occurs in strong trends where RSI doesn't reach traditional extremes. In uptrends, RSI oscillates 40-80 (never touches 30). In downtrends, RSI oscillates 20-60 (never touches 70). To use it: identify the recent RSI range over 20-30 periods. If lowest RSI has been 42, that's the 'oversold' level for this trend. Buy when RSI reaches the bottom of its recent range, not necessarily 30.

Create a 'confluence checklist': (1) Identify key S/R levels from the daily chart, (2) Mark round numbers ($3,500, $3,600, etc.), (3) Note SHFE-implied equivalent levels. When price approaches any key level, check RSI. If RSI is also at an extreme, you have confluence. Enter only on confluence setups initially - they have much higher success rates. Without confluence (RSI extreme at a random price), be more cautious or use a smaller position.

Key components: (1) Clear signal rules - RSI crossing 30/70 with close confirmation, (2) Trend filter - 50 EMA alignment, (3) Volume filter - signal candle > 80% average volume, (4) Position sizing - 2% risk / ATR-based stop, (5) Exit rules - opposite RSI zone or ATR-based stop/trail. Backtest 3+ years including different market conditions. Target: win rate >50%, profit factor >1.5, max drawdown <12%. Use walk-forward optimization to prevent curve fitting.

Most effective modifications: (1) Smoothed RSI using EMA instead of SMA - reduces whipsaws, (2) Adaptive period - shorter (10) in high volatility, longer (18) in low volatility, (3) Connors RSI combining standard RSI with streak and percentile rank. Backtest each modification independently. For Aluminium specifically, the smoothed RSI typically shows 10-15% improvement in profit factor over standard RSI.

Scaling framework: Entry 50% when RSI crosses 30, add 25% when RSI crosses 40, add 25% when RSI crosses 50. For exits: take 30% at RSI 55, 30% at RSI 65, trail remaining 40% until RSI drops below 55. Key rules: never add to losing positions, each add should have RSI progressing in your favor, reduce adds if momentum stalls (RSI flattening). This method can improve returns 30-50% over fixed position approaches.

Create a multi-asset RSI dashboard: LME Aluminium, SHFE Aluminium, Dollar Index, Copper. Best long setup: LME oversold + SHFE oversold + DXY overbought + Copper not extremely bearish. Monitor each instrument's RSI alongside Aluminium. When 3+ factors align, increase position size. When only the LME signal exists without intermarket confirmation, reduce size or skip. This approach significantly improves signal quality.

Key metrics: (1) Win rate by RSI level (are 25 readings more profitable than 29?), (2) Profit factor by session (Asian vs European vs US), (3) Average hold time for winners vs losers, (4) Performance with/without trend filter, (5) MAE/MFE analysis for stop optimization, (6) Win rate at confluence vs non-confluence levels, (7) Performance in trending vs ranging markets (use ADX to classify). Review monthly, adjust parameters quarterly based on data.