MACD Momentum

These defaults, established by Gerald Appel, roughly correspond to meaningful trading periods: 12 days ≈ 2.5 trading weeks, 26 days ≈ 1 trading month (about 5.5 weeks), and 9 days for the signal line provides good smoothing. They've become standard because many traders use them, creating a self-fulfilling element. They're not magic numbers - they're a reasonable balance between responsiveness and noise reduction that has worked well historically.

MACD provides additional information that simple MA crossovers don't. The histogram shows momentum strength and whether it's accelerating or decelerating. MACD can show divergences that warn of trend exhaustion. However, MACD is essentially a derivative of MA crossovers (it IS the difference between two EMAs), so they often give similar signals. MACD is better for momentum analysis; simple MAs are simpler to understand and use.

On daily charts with standard settings, expect roughly 2-4 signals per month per instrument - though this varies with market conditions. Trending markets produce fewer, cleaner signals. Choppy markets produce more signals but with more whipsaws. If trading a portfolio of 10 instruments, you might see 20-40 signals monthly across the portfolio, giving you plenty of opportunities to be selective.

Yes, MACD works on any timeframe, but you may want to adjust settings. For day trading on 5-minute charts, faster settings like 8/17/9 or even 5/13/6 help catch moves earlier. However, shorter timeframes generate more noise and false signals. Many day traders prefer MACD on 15-minute or hourly charts for signal generation while using lower timeframes for entry timing.

Not necessarily immediately. Shrinking histogram warns momentum is fading, but it doesn't mean the move is over. It's a signal to tighten your stop or watch for the actual crossover. Some traders exit half on histogram fade and the rest on crossover. Others just use it to mentally prepare for exit rather than acting immediately. The full crossover remains the primary exit signal for most MACD traders.

Reliable divergences typically: (1) Occur after extended trends, not early in moves, (2) Show clear, well-defined peaks/troughs in both price and MACD, (3) Are confirmed by other factors like volume divergence or price pattern completion, (4) Occur at significant price levels (major support/resistance). Unreliable divergences often: Occur in choppy, trendless markets, show ambiguous peaks, or appear after only minor price moves. Divergences work best as warnings to tighten stops, not as primary entry signals.

Use the zero line filter when: Trading position/swing style where you want higher conviction, the market has clear trends, and you can afford to miss some moves. Ignore the zero line when: Day trading where you need more signals, the market is range-bound (all signals would be filtered), or you're experienced at managing counter-trend trades. A middle approach: Use zero line for position sizing - full size when aligned, half size when not.

Use them for different purposes rather than requiring both to agree. MACD for trend/momentum direction and RSI for overbought/oversold extremes. Example: MACD bullish crossover (entry trigger) + RSI not overbought (<70) = good entry. MACD bullish crossover + RSI very overbought (>80) = wait for RSI pullback before entering. Or use RSI divergence to anticipate MACD crossovers. Don't require perfect alignment - you'll miss too many trades.

Rank signals by quality and take the best ones. Ranking criteria: (1) Higher timeframe alignment, (2) Histogram strength (taller bars = more momentum), (3) Volume confirmation, (4) Price action confirmation, (5) Distance from zero line. If 5 signals appear and you can only take 3, select the three with the best combined scores. This ensures capital goes to highest-probability setups.

If a position is flat (not moving) for an extended period, it's using capital that could be deployed elsewhere. Rules of thumb: If no meaningful progress after 10-15 bars (about 2-3 weeks on daily), reassess. If histogram is flat and not growing, momentum isn't there. Consider a time-based exit if position hasn't moved significantly. However, don't be too quick - giving trades time to work is part of the system. Balance patience with capital efficiency.

Monitor signal effectiveness rolling forward. If win rate drops significantly below historical norms for 2-3 months, regime may have changed. Key indicators: (1) VIX level shifts (sustained move above 25 or below 15), (2) ATR percentile changes, (3) Market transitioning from trending to ranging. Implement systematic regime detection: measure current volatility vs. historical percentiles and adjust MACD periods proportionally. Slower settings in high-vol, faster in low-vol.

For individual stocks: Avoid new entries 5-7 days before earnings - binary risk negates technical signals. For existing positions: Either exit before earnings or hedge with options. After earnings: Wait 1-2 days for dust to settle before acting on new MACD signals - post-earnings price discovery often creates false crossovers. For index trading (SPY, QQQ): Be aware of heavy earnings periods but index diversification reduces single-stock risk. May want to reduce overall position sizes during peak earnings season.

Essential metrics: (1) Win rate by signal type (bullish vs bearish, above/below zero), (2) Average R-multiple (risk-adjusted return), (3) Profit factor (gross wins/gross losses), (4) Maximum drawdown and recovery time, (5) Sharpe/Sortino ratios, (6) Signal-to-noise ratio (profitable signals/total signals), (7) Execution quality (slippage vs. expected), (8) Strategy decay analysis (performance over time). Compare metrics across market regimes to understand when system works best/worst.

Scan sector ETFs (XLK, XLF, XLE, etc.) weekly for MACD signals. When a sector shows bullish crossover above zero with strong histogram, it indicates money flowing into that sector. Rotate portfolio toward sectors with strongest MACD readings and away from weakest. Compare sector MACD to SPY MACD - sectors outperforming SPY's MACD show relative strength. This creates a systematic sector rotation approach based on momentum rather than fundamentals.

Key practices: (1) Use round numbers with economic rationale (week, month periods) rather than arbitrary optimized values, (2) Test sensitivity - if 12/26/9 works but 11/25/8 fails badly, the edge is fragile, (3) Walk-forward optimization - optimize on period 1, test on period 2, optimize on periods 1-2, test on period 3, etc., (4) Keep parameters constant for extended periods (quarters/years), not adapting to recent performance, (5) Accept that 'good enough' beats 'perfect' - marginal optimization often just captures noise, (6) Cross-validate across multiple instruments.