EMA vs SMA Trend Following Playbook

Execution snapshot

Market regime
Strong directional moves with orderly pullbacks.

Primary pairs
BTC/ETH majors, liquid alt majors.

Typical bands
first_band ema 21–34 vs second_band sma 55–89.

Signal rule
Long when EMA closes above SMA; short when EMA closes below.

Why this works

The EMA weights recent candles more heavily, letting it respond quickly to fresh momentum. The SMA averages a wider sample, smoothing the macro bias. When the fast line overtakes the slow line on a closed bar, we have evidence that the most recent trend impulse is dominating prior price memory.

Because crypto majors trend for extended stretches, the EMA/SMA pairing keeps you positioned for the move without requiring discretionary confirmation on every swing.

Configuring in QuantRelay

first_band: ema with length between 21 and 34. Use 21 on 1h charts, 34 on 4h when you need extra smoothing.
second_band: sma with length 55–89. Higher values filter sideways chop at the cost of slower reactions.
Keep offset = 0 for both to avoid look-ahead bias.
For volatile altcoins, add a daily_profit filter in the main index to surface the most resilient performers.

Playbook

Entry

Wait for a closed candle where the EMA crosses the SMA in the desired direction.
Confirm the higher timeframe (e.g., 4h if trading 1h signals) is aligned to avoid counter-trend traps.
Size positions based on the distance between EMA and SMA — wider spread signals stronger trend conviction.

Exit

Either trail stops below the SMA (for longs) or use ATR-based risk bands.
Flip to the opposing signal once the EMA crosses back through the SMA, but reduce size after extended runs where mean reversion is likely.

Risk controls

Avoid trading immediately around major macro events; EMAs whip violently on liquidations.
Layer in volume filters — declining volume on upward crosses hints at false breakouts.
Use partial profit-taking at predefined RR targets (e.g., 1.5R) to lock gains before consolidation.

Going further

Combine with supertrend as a higher-timeframe confirmation to cut whipsaws.
Experiment with hma or jma as the fast line for smoother responses.
Backtest length variants directly in QuantRelay using the Copy → v2 action to clone into your sandbox table.