Edge Library — Filter Effect Reports

Platform-wide aggregated analysis: how each Pro+ entry filter affects strategy CAGR (baseline vs filtered), asset-equal-weighted — per-asset medians over param-deduplicated runs, then the median across assets. delta_cagr is the median of per-asset deltas over MATCHED assets (present in both arms) and therefore usually differs from filtered_cagr − baseline_cagr; pairs_matched/pairs_filtered plus the baseline pairs count declare the basis. Verdicts come from the effect's 90% paired-bootstrap interval, not the point estimate: helps (whole interval > +1pp), hurts (whole interval < −1pp), neutral (whole interval inside ±1pp), insufficient_evidence (enough runs, but the interval straddles the line), insufficient_data (fewer than 30 runs per arm or fewer than 10 matched assets). Interval bounds are returned as delta_ci_low / delta_ci_high. Response is GROUPED by strategy: envelope fields (market, computed_at, n_trials) once, per strategy one baseline block plus filter cells; zero-run cells are folded into filters_without_data; truncated flags a partial page. Filters evaluated in isolation (no stacking). Net values are median CAGR after per-side trading costs (verdict/delta stay gross). Below the 30-run gate, derived fields (delta_*, dsr, dsr_pass) are null rather than served; every gated null carries its reason (dsr_pass_reason, baseline_net_reason / filtered_net_reason); the envelope evidence block declares the gate's referent (gate_applies_to) and threshold machine-readably. Values are server-rounded (2dp percent, 4dp Sharpe). [Free tier]