Five pillars, 20 points each, scored on piecewise-linear curves. Everything is
computed from live data at page load — nothing here is hardcoded.
1. VIX Level — is there premium worth selling?
Peaks around VIX 22–28. Scores low under 13 (nothing to sell) and falls off again
above 35, because a VIX that high usually means short-premium positions are getting run over.
2. VIX Trend — crushing or spiking?
Current VIX vs its 10-day average. Falling volatility is IV crush working in your
favour (max score); a sharp spike scores near zero.
3. IV Percentile — is implied vol rich?
For each name, where today's at-the-money IV sits within that stock's own past year of
30-day
realized volatility, averaged across the basket. True IV rank needs a year of
stored IV history, which no free API gives — so this uses realized vol as the yardstick.
The app now records IV daily in
cache/premium_history.json, so a real
IV rank becomes possible once a year of scans has accumulated.
4. SPY Trend — direction risk
Position vs the 50 and 200-day moving averages, adjusted by the 5-day change. Selling
puts into an uptrend scores well; a falling knife does not.
5. Premium Yield — what you actually get paid
Mid price ÷ strike on the ~30-delta put, annualized, averaged across the basket.
Data sources
Yahoo Finance chart API (^VIX, SPY, per-ticker history) and the CBOE delayed-quotes
API (full option chains with IV, delta, bid/ask, open interest). Both are free and need
no API key. Option quotes are
15-minute delayed and go stale outside
market hours. Results are cached for 3 hours; Rescan forces a refresh.
Bands
0–39 poor · 40–59 mixed · 60–79 favorable · 80–100 prime.
Full methodology — exact thresholds, data limits, and what this engine deliberately does not do →