A daily-updated defensive regime monitor for the S&P 500: a five-component composite score (0–5) drives a defensive allocation ladder, with the full indicator set, episode statistics and strategy comparison behind it.
Live: https://phuazz.github.io/equity-defense-dashboard/
Personal research artefact. Not investment advice, not an offer, and not affiliated with any regulated fund. All returns shown are simulated; there is no live track record. The page footer carries the full disclosure.
One point per active signal, evaluated daily:
| # | Signal | Trigger |
|---|---|---|
| 1 | Stress count | Rolling 8-day count of S&P 500 members down ≥7%, scaled to index size (Batnik's "blowup" count) |
| 2 | 200-day MA | SPX below its 200-day simple moving average |
| 3 | 12-month momentum | SPY trailing 12-month total return negative (Antonacci absolute momentum) |
| 4 | 10-month SMA | SPX below its 10-month SMA at the monthly read (Faber 2007) |
| 5 | VIX term structure | VIX above VIX3M (inverted term structure) |
Allocation ladder by score: 0–1 → 100% SPY · 2 → 50% SPY / 50% defensive · 3 → 25% SPY / 75% defensive · 4–5 → fully defensive. Defensive asset: SHY (1–3y US Treasuries) primary, IEF (7–10y) shown for comparison. Episode statistics are measured from the first day a score level is reached, with forward returns from T+1.
Tabs: Monitor (regime dial, allocation, what-to-watch) · Indicators (each signal against its trigger) · Performance (forward-return and drawdown distributions by score level, episode analysis) · Methodology.
- Adjusted close throughout (dividends and splits reinvested — total return).
- Universe: 90 S&P 500 representative members, counts scaled by
500 / available(currently ×5.56); plus SPY, SHY, IEF and VIX/VIX3M. - Window: 1998-01-05 onward; the masthead shows the live end date.
- Providers: Yahoo Finance (default) or Alpha Vantage behind a common
DataProviderinterface, with an incremental committed cache (data/cache.json) so each daily run fetches only the delta.
.github/workflows/update.yml daily cron 01:00 UTC
scripts/data_providers.py provider abstraction + cache
scripts/compute_signals.py pure-Python signal engine
scripts/pipeline.py fetch → compute → render
data/signals.json computed signals + metadata
template.html source page (edit this)
docs/index.html built page (never edit — GitHub Pages serves /docs)
Local build: pip install -r requirements.txt then python scripts/pipeline.py
(first run fetches full history; later runs use the cache). --no-cache
forces a refetch; --provider alphavantage switches provider (needs
ALPHA_VANTAGE_KEY).
- Live and refreshing daily — the Actions cron commits "📊 Update signals" each session day; data end date renders in the masthead.
- Cross-dashboard audit 2026-08-13 (
C:\dev\studies\2026-08-13_dashboard-audit.md): disclaimer footer added, sub-11px type raised to the mobile floor, README rewritten (this file — the previous one described the pre-rename "blowup-signal" project with a placeholder URL). - Related project: breadth-thrust-signal — the bullish mirror (risk-on conviction meter).
- Batnik, M. — "The Blowup Signal" (The Compound)
- Faber, M. (2007) — "A Quantitative Approach to Tactical Asset Allocation"
- Antonacci, G. (2014) — "Dual Momentum Investing"
- Keller, W. & Keuning, J. (2018) — "Defensive Asset Allocation"
- Zweig, M. (1986) — Breadth Thrust indicator
- Siegel, J. (2006) — "Stocks for the Long Run"
MIT licence.
Last updated: 2026-08-13