Aluminum RSI Strategy

RSI works well for aluminum 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, aluminum's steadier price action means oversold and overbought readings more reliably lead to reversals. The moderate daily moves allow RSI signals to play out without excessive noise.

No - there is no domestic Canadian aluminum futures contract. The Montreal Exchange lists equity, index and interest-rate derivatives, not base metals. Canadian traders access aluminum on COMEX (the ALI contract, 25 metric tons, priced in USD) through a CIRO-regulated broker that offers US futures market access, such as Interactive Brokers Canada. There is no mini contract, so the smallest position is one full 25-tonne contract, and your P&L is in USD - meaning you also carry a USD/CAD currency exposure.

RSI works alone but combining it with filters improves results significantly. The most effective addition is a 50 EMA trend filter - only take buy signals above the 50 EMA, sell signals below. This simple filter can improve win rate from around 50% to roughly 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 the higher timeframe RSI for context - daily RSI above 50 means a 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) the higher timeframe RSI supports the trade direction, (3) volume confirms the move, and (4) LME aluminum is moving in the same direction. Avoid RSI signals during thin overnight (non-US) liquidity, at random price levels, or against the major trend.

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 the direction of the higher timeframe trend for best results, and (4) set a 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 a shorter RSI period (10-12) for faster signals or raise overbought/oversold levels to 75/25 to filter noise. In low volatility: the standard 14-period works well, and 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, and for a Canadian trader it is even more central because there is no domestic contract to reference. COMEX aluminum tracks the LME closely. If COMEX RSI is oversold but the LME is trending down strongly, the bounce may be limited. The best signals occur when both COMEX and LME RSI are oversold and showing signs of reversal. Signals during thin, non-US hours without LME 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 the 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 ($2,500, $2,600, $2,700), and (3) note the equivalent LME 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. Remember the USD-denominated P&L means a Canadian trader's realised CAD return also depends on USD/CAD.

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 aluminum specifically, the smoothed RSI typically shows a 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 the 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). Note the COMEX constraint: with no mini contract, scaling requires holding several full 25-tonne ALI contracts, so this method needs a larger account.

Create a multi-asset RSI dashboard: COMEX aluminum, LME aluminum, Dollar Index, copper. Best long setup: COMEX oversold + LME oversold + DXY overbought + copper not extremely bearish. Also track the CME Midwest Premium and Section 232 tariff headlines, since these move the North American premium independently of the global benchmark. When 3+ factors align, increase position size. When only the COMEX signal exists without intermarket confirmation, reduce size or skip.

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