VIX DMA Edge: variant parity + data coverage audit (qm-a3a.1)
Rejected qm-a3a.1 ·
Verdict
No implementation bugs found. Probability / Kelly / SH variant params change signal exits and position sizing (weight series differ). Headline metrics still look nearly identical because Kelly defensive stakes are economically small and edge days are sparse — genuinely similar, not no-op code paths. Parent qm-a3a NO-SHIP stands; monthly SPY corr ≈ 0.99 rules out a diversifier-only ship
Provenance
Equity vs S&P (aligned, rebased to 1.0)
Aligned window, both rebased to 1.0 at start — Aligned window, both rebased to 1.0 at start for fair comparison · S&P dashed, experiment solid. X = Date, Y = Equity (rebased to 1.0).
Signal driver & thresholds (x-aligned)
VIX DMA distribution
Probability exits, Kelly sizing
Stats
| CAGR | 2.90% | Sharpe | 59.80% |
|---|---|---|---|
| Sortino | 87.21% | maxDD | -19.30% |
Disposition
# VIX DMA Edge: variant parity + data coverage audit (qm-a3a.1) ## Verdict **No implementation bugs found.** Probability / Kelly / SH variant params change signal exits and position sizing (weight series differ). Headline metrics still look nearly identical because Kelly defensive stakes are economically small and edge days are sparse — **genuinely similar**, not no-op code paths. Parent **qm-a3a NO-SHIP** stands; monthly SPY corr ≈ 0.99 rules out a diversifier-only ship rationale. ## Weight differentiation (not no-ops) | Pair | Left | Right | n_diff / n_common | max \|Δw\| | |---|---|---|
Follow-on work
- Follow-up qm-a3a.1
- Follow-up qm-a3a
- Follow-up qm-a3a.1.json
Errors & data gaps
| Ticker | First | Last | n_bars | Source | Notes | |---|---|---|---:|---|---| | `SPY` | 1993-01-29 | 2026-07-24 | 8428 | Yahoo via market.sqlite ensure_ohlcv | bday_gap_count is mostly US holidays vs pandas bday; max consecutive bar gap 7d; holes>10d=0. | | `^VIX` | 1990-01-02 | 2026-08-04 | 9215 | Yahoo via market.sqlite ensure_ohlcv | bday_gap_count is mostly US holidays vs pandas bday; max consecutive bar gap 7d; holes>10d=0. | | `SH` | 2006-06-21 | 2026-08-06 | 5063 | Yahoo via market.sqlite ensure_ohlcv | Live SH IPO; used only for short_on_upper live-panel variants. Pre-IPO history uses
Follow-up
gh issue create --title "Follow-up VIX DMA Edge variant parity data coverage audit qm-a3a1" --body "Follow-up to qm-a3a.1: Verdict" --label "priority:p2" --label "pipeline:research" --label "agent:ready"