Reading Options Edge — Weekly IV Anomaly Scan
Options Edge runs every Sunday across a locked 25-ticker universe (indexes, mega-cap tech, semis, high-IV retail, sector ETFs). For each ticker it computes four volatility-surface metrics and z-scores them against the ticker's own 1-year history. Anomalies — defined as |z| ≥ 2.0 — get surfaced as ranked cards with suggested strategies and concrete strike chips. The math is deterministic; the routine-written prose adds market context.
What you're looking at
Every Sunday afternoon, the Options Edge routine pulls the live options chain for all 25 watchlist tickers, computes four IV-surface metrics per ticker, and z-scores each against the ticker's own 1-year history stored in the iv_snapshots table. Anomalies (|z| ≥ 2.0) are ranked by absolute z-score and published as cards on this page, with a routine-written prose summary up top and a deterministic strategy + strike suggestion per card.
The four metrics, in plain English
1. ATM IV rank
Where today's 30-day at-the-money implied vol sits within the ticker's last 252 days of ATM IV. Rank 95 = top 5% of the 1-year range (vol is historically expensive). Rank 5 = bottom 5% (historically cheap). High rank favors selling premium structures; low rank favors buying gamma.
2. 25-delta skew z-score
The spread between the 25-delta put IV and the 25-delta call IV, z-scored against the ticker's 1-year norm. Positive skew (puts richer than calls) is normal in equity options; what matters is whether the spread is at its NORMAL level or stretched. High z = puts unusually rich → risk-reversal (sell put, buy call) captures the mean reversion. Low z = puts unusually cheap → reverse risk-reversal.
3. Term-structure z-score
The 60-day ATM IV minus the 30-day ATM IV, z-scored. Normal markets have positive term (longer-dated vol > near-dated). When this widens unusually high, near-dated vol is cheap relative to back month — calendar spreads work. When it flips negative (inverted), front-month vol is bid for an event — buy back-month, let front bleed.
4. IV/HV ratio z-score
Implied vol divided by realized vol, z-scored. This is the "variance risk premium" — how much extra the market is paying for protection vs what the underlying is actually moving. High z = fat premium, short-vol pays. Low z = realized is hotter than implied, long-vol pays.
How to use the page
- Read the green Anomalies box first.The 2-3 most extreme picks of the week with the routine's context.
- Scan the ranked list below. Cards are ordered by |z| descending — biggest deviations first. Each card has everything you need to evaluate: surface values, suggested structure, concrete strikes.
- Cross-check with the prose summary.Catalysts, earnings within 21 days, Fed meetings, OPEX — the routine flags anything that could explain (or undermine) the statistical signal. A high-z anomaly on a name with earnings next week isn't mispricing; it's event premium.
- Verify strikes on your broker. The chip strikes are computed from a model — listed strikes might be a tick off.
What this is NOT
Options Edge is a statistical anomaly screener, not a directional forecast. A +2σ ATM IV rank doesn't mean the stock will move less; it means vol is statistically expensive, and short-vol positions are FAVORED — not guaranteed. Black swans break the pattern. Size accordingly.
Frequently asked questions
What are the four metrics Options Edge scans?
(1) ATM IV rank — current 30-day at-the-money IV as a percentile of its 1-year range. (2) Skew z-score — the 25-delta put IV minus 25-delta call IV, z-scored against 1y norm. High = puts unusually rich vs calls. (3) Term structure z-score — 60-day ATM IV minus 30-day ATM IV. High = unusual contango; low = inversion (front-month event premium). (4) IV/HV ratio z-score — implied vol divided by realized vol. High = fat variance risk premium (vol is expensive vs what's actually printing).
Why z-score against 1-year history?
Raw IV numbers don't tell you anything alone — TSLA at 60% IV could be expensive or cheap depending on its baseline. Z-scoring against the ticker's own 252-day history makes 'expensive vs cheap' meaningful: z=+2 means the current value is two standard deviations above its 1-year mean, which historically marks the kind of extreme that mean-reverts. The threshold |z| ≥ 2.0 surfaces only statistically meaningful deviations.
What does the green 'Anomalies' box at the top show?
The 2-3 most extreme anomalies of the week, lifted out of the routine's prose summary into a hero box for scan-and-go reading. Each line names the ticker, the anomalous metric, the z-score, and the routine's color around why it matters. The same picks also appear lower in the ranked-anomalies list with full surface data and suggested strikes.
How do I read a ranked anomaly card?
Top row: ticker · metric chip · direction chip (Stretched high / Stretched low) · z-score and percentile. Strategy line: the routine's suggested trade structure. Strike chips: emerald = buy legs, rose = sell legs, with concrete strikes computed via delta-target inversion and snapped to the listed grid. Surface mini-table: underlying spot, ATM IV 30d, 25Δ put IV, 25Δ call IV, HV 30d — the raw context. Read top to bottom for the call.
How are the suggested strikes computed?
Strikes use the standard log-normal delta-target approximation: K(δ) ≈ S · exp(±N⁻¹(δ) · σ · √T), with N⁻¹(0.25) = 0.6745 for 25Δ legs and N⁻¹(0.10) = 1.2816 for 10Δ wings. T = 30/365. Output is snapped to the nearest listed-options grid (typically $5 above $200, $1 below). They're approximations — meant as a starting point, not a quote. The actual chain might have slightly different listed strikes; verify on your broker before sending an order.
What's a 'Notable extreme' callout at the top?
When any single anomaly has |z| > 3.5 (roughly once per quarter per metric per name), the routine flags it as Notable Extreme. That's tail-of-distribution territory — historically the most reliable mean-reversion setup. Doesn't guarantee anything, but it's the highest-conviction signal the scanner produces.
How should I size and time these trades?
Size small. These are statistical mean-reversion trades, not directional certainties — a +2σ anomaly can become a +3σ before reverting. Timing: most short-vol structures (iron condor on rich IV) work best held to expiration or rolled at 50% max profit. Long-vol structures (straddle on cheap IV) need a catalyst — earnings, Fed, OPEX — to monetize the vega expansion. The 'Risks & caveats' section in the routine's summary flags upcoming events that could change the picture.
Why only 25 tickers? Can the universe expand?
The universe is locked because every name requires 1 year of IV history (the iv_snapshots backfill) to z-score against. Expanding means running the backfill for the new ticker first, which is a one-time job per name. The current 25 covers indexes (SPY/QQQ/IWM), mega-cap tech, semis, high-IV retail, and sector ETFs — enough breadth to surface meaningful anomalies every week without being noisy.
What if NO anomalies cleared the bar?
It still publishes — 'No anomalies cleared the |z| ≥ 2.0 threshold this scan. The volatility surface across the universe is sitting within its 1-year norms.' The absence of edge is itself information: a calm regime where short-vol systematic strategies work fine but there's no specific edge worth chasing.
How does Options Edge fit with Unusual Activity and GEX?
Three layers of the same picture. Options Edge = the SURFACE (where is vol mispriced statistically). Unusual Activity = the FLOW (who's positioning, and how aggressively). GEX = the HEDGING (what dealers will be forced to do as price moves). A trade with all three aligned — say, low IV rank + heavy bullish call buying + positive dealer gamma at the strike — is much higher conviction than any one signal alone.