cboe-data

CBOE Volatility Indices Reader

Pull and interpret CBOE indices for market-vol regime context.

---

Step 1: Pull the Index Snapshot

Via yfinance (free, delayed) or CBOE direct API.

import yfinance as yf
INDICES = {
    'VIX':    '^VIX',    # 30-day expected SPX vol
    'VVIX':   '^VVIX',   # vol of VIX
    'SKEW':   '^SKEW',   # tail-risk premium
    'VIX9D':  '^VIX9D',  # 9-day expected vol
    'VIX3M':  '^VIX3M',  # 3-month expected vol
    'VIX6M':  '^VIX6M',  # 6-month expected vol
    'PCRATIO': None,     # CBOE publishes daily; not on yfinance
}
def snapshot():
    out = {}
    for name, sym in INDICES.items():
        if sym:
            t = yf.Ticker(sym)
            out[name] = t.fast_info['lastPrice']
    return out

If the snapshot fails for any index, list which ones — don't paper over with NaN.

---

Step 2: Interpret Levels

Index	Calm	Normal	Elevated	Stressed
VIX	< 12	12–18	18–25	> 25
VVIX	< 80	80–100	100–120	> 120
SKEW	< 120	120–135	135–150	> 150
PCRATIO (equity)	< 0.6 (greed)	0.6–1.0	1.0–1.3	> 1.3 (fear)

Levels alone aren't enough — show 1y percentile too.

---

Step 3: Term-Structure Ratios

ratio_9d_30d = VIX9D / VIX
ratio_30d_3m = VIX / VIX3M
ratio_30d_6m = VIX / VIX6M

Ratio	< 1	> 1
VIX9D / VIX	Curve normal (contango)	Frontend stress (backwardation)
VIX / VIX3M	Front below back (calm)	Front above back (stress)

Sustained VIX > VIX3M for several days is the canonical "buy SPX" signal at extremes (mean reversion of curve). Sustained VIX9D / VIX > 1 precedes vol expansion more often than not.

---

Step 4: Cross-Index Signals

Signal	Pattern	Read
Complacency	VIX low, SKEW high, P/C low	Tail risk priced but spot vol cheap — gap risk
Fear bid	VIX high, VVIX high, SKEW high	Generalized stress, vol-of-vol elevated
Vol crush ahead	VIX9D / VIX > 1.1, no known event	Mean reversion likely; consider selling front-end vol with a stop
Earnings season	VIX > VIX3M but only on event-heavy weeks	Don't read as market-wide stress

---

Step 5: Respond to User

1. Snapshot table (index, level, 1y percentile, classification).
2. Term-structure ratios.
3. One-line regime call (complacent / normal / stressed / event-driven).
4. Trade ideas only if a strong signal — be specific about which expiry / which underlying. No generic "consider buying puts".

Recommend vol-term-structure to drill into a specific name, or vol-skew for per-name skew.

---

Caveats

• VIX is 30-day forward, model-derived from SPX option mids — it's not "what vol is now", it's "what option prices imply about the next 30 days".
• VVIX is vol of VIX; spikes when traders gamma-hedge VIX options.
• SKEW is non-trivial to interpret in absolute terms — its 1y percentile is more useful than the raw number.
• yfinance VIX data can be ~15min delayed. For real-time use CBOE's paid feed.