Does ATR Expansion help EMA Trend Bias?
Crypto · 92 backtest runs · As of 15 Jun 2026 · Methodology
HurtsDSR ✗ not confirmed
This filter reduces median CAGR by more than 1 percentage point — either more false signals or too many blocked trades.
CAGR Comparison
Baseline CAGR
-2.0%
no filter
Filtered CAGR
-4.8%
ATR Expansion
Δ CAGR
-2.70%
Sample
92
runs
Risk-Adjusted Performance
Baseline Sharpe
0.091
median
Filtered Sharpe
-0.081
Δ Sharpe
-0.17
DSR
0.0%
not confirmed
How does this filter work?
Signals only during ATR expansion (trend onset)
DSR Methodology
DSR = PSR(SR̂₀) per Bailey & López de Prado (2014). SR̂₀ = 0.5852 (expected best Sharpe from 7 random filter trials).
Testing multiple filters on the same data increases the chance of finding a good result by luck. DSR measures whether this result clears the multiple-testing threshold. Pass threshold: DSR ≥ 90%. Current: 0.0%.
Full analysis in Edge Library
Per-asset breakdown, trade-level data, Sharpe comparison — available after sign-in.
← All edge reportsComputed 15 Jun 2026 · Methodology