GEXLOG Methodology
How we measure dealer gamma exposure on the S&P 500 — and what the numbers mean for session character.
What We Measure
GEXLOG analyzes dealer gamma exposure (GEX) on the S&P 500 to estimate how market makers will be forced to hedge — and whether that hedging will dampen or amplify price movement on any given session.
Every metric on this dashboard is pre-computed from options chain data before the market opens. Nothing is inferred by AI. The models receive pre-calculated facts and structure narratives around them.
Gamma Exposure (GEX)
Nearest-Expiration Only
GEXLOG calculates GEX using the nearest expiration (typically 0–1 DTE) of SPXW weekly options. This isolates the contracts most likely to drive same-day dealer hedging flows.
Many third-party GEX providers aggregate gamma across all expirations. Both approaches are valid — they answer different questions:
| GEXLOG (Nearest Expiration) | Multi-Expiration Aggregate | |
|---|---|---|
| Focus | Same-day dealer hedging pressure | Structural positioning across term structure |
| Flip sensitivity | May show no flip when short-dated puts dominate | More likely to show a flip from longer-dated call OI |
| Best for | Intraday regime and session character | Broader market structure and multi-day thesis |
If GEXLOG shows no GEX flip level and another provider does, this is the most common reason. In deeply negative gamma environments, cumulative GEX across the nearest expiration may never cross zero — put GEX overwhelms call GEX at every strike. This is reported accurately as "no flip level exists" rather than manufacturing an approximation.
GEX Formula
GEX per strike = gamma × open_interest × 100 × (SPX² / 100)
- Calls contribute positive GEX (dealers short calls → long delta → sell rallies)
- Puts contribute negative GEX (dealers short puts → short delta → buy dips)
- Net GEX = sum of all call GEX + sum of all put GEX
The gamma values come from the Tradier options chain API with greeks enabled. This is GEX-weighted open interest — not raw OI concentration.
Regime
- Net GEX > 0 → Positive gamma: Dealers hedge against moves. Volatility is suppressed. Sessions tend to be range-bound and mean-reverting. Favors income strategies.
- Net GEX < 0 → Negative gamma: Dealers hedge with the trend. Moves are amplified in both directions. Sessions tend to be trending and momentum-driven. Favors directional, defined-risk strategies.
Data Timing
GEX is calculated pre-market (~6:45 AM ET) using settled end-of-day OI and greeks from the prior session. It is recalculated in the evening (~7:45 PM ET) targeting the next session's chain.
SPX index options data via Tradier carries a mandatory 15-minute OPRA delay, which is immaterial for batch pre-market processing but would make intraday recalculation unreliable. GEXLOG does not attempt intraday GEX updates.
Key Levels
Put Wall & Call Wall
The Put Wall is the strike with the largest absolute put GEX below spot price. The Call Wall is the strike with the largest call GEX above spot price.
Both are bounded by a search window of 3× the daily expected move from the current SPX price, hard-capped at 5% of spot. This prevents distant, illiquid strikes from appearing as walls.
Wall selection uses a tiered threshold: a strike must concentrate ≥5% of total side GEX (strict) or ≥1.5% (soft fallback). If no strike qualifies, the wall falls back to the nearest round number (multiples of 50).
Walls and the GEX flip are independent fields. A null flip does not mean walls are unavailable, and vice versa.
GEX Flip
The strike where cumulative GEX (walking from lowest to highest strike) transitions from negative to non-negative. This marks the boundary between negative and positive gamma territory.
When cumulative GEX never crosses zero — common in deeply negative gamma environments — the flip is reported as null. No approximation is manufactured.
Pivot Levels (R1 / R2 / S1 / S2)
Classic floor trader pivot points calculated from the prior session's SPX high, low, and close:
Pivot = (High + Low + Close) / 3
R1 = (2 × Pivot) - Low
R2 = Pivot + (High - Low)
S1 = (2 × Pivot) - High
S2 = Pivot - (High - Low)
Pivots are entirely independent of GEX-derived levels. They are pure price math from prior session OHLC. When pivots and walls converge at similar levels, that confluence is analytically meaningful. When they diverge, each system is telling a different story.
If the market gaps above R1, R1 becomes a support reference below price. This is normal behavior — pivots are static.
Expected Move
Calculated using the Rule of 16:
Daily Vol = VIX / 15.87
Expected % = Daily Vol × √DTE
Expected Move = SPX Price × (Expected % / 100)
DTE is floored at 1 to prevent mathematical collapse on 0DTE chains. The expected move is reported as a point value (e.g., ±106 points), not a percentage.
- Price within ±1 EM = normal session
- Price exceeding ±1 EM = expansion / high-volatility session
VIX Regime Thresholds
| VIX Level | Classification |
|---|---|
| Below 14 | Low / Complacent |
| 14–20 | Normal |
| 20–25 | Elevated |
| Above 25 | High Fear |
These thresholds are pre-computed and injected into AI prompts as facts. The AI does not interpret "high VIX" subjectively.
ES Futures vs. SPX Cash
GEXLOG references both instruments and they are not interchangeable:
- SPX (S&P 500 Index): Cash index. Regular hours only (9:30 AM–4:00 PM ET). All options-derived levels, pivots, and expected move are anchored to SPX.
- ES (E-mini S&P 500 Futures): Futures contract. Trades nearly 24 hours. Used as a pre-market proxy for where SPX will open.
In pre-market content, GEXLOG presents distances to key levels from both ES futures (current state) and SPX prior close (gap-fade fallback).
Gap Analysis
The overnight gap is calculated as the percentage difference between the current ES futures price and the prior SPX close.
| Gap Size | Classification |
|---|---|
| < 0.2% | Small — fade candidate |
| 0.2%–0.5% | Moderate |
| > 0.5% | Large — follow-through candidate |
Calendar Risk
Economic events are scored 1–5 based on historical market impact:
| Score | Events |
|---|---|
| 5 | FOMC, CPI, Non-Farm Payrolls |
| 4 | PPI, GDP, PCE |
| 3 | Jobless Claims, Retail Sales, Major Earnings |
| 2 | Fed Speakers |
| 1 | Other |
The daily calendar risk is the maximum score among all events for that session.
AI Narratives
Morning reports, evening recaps, and YouTube briefs are generated by Claude and Gemini. Every factual claim in the narrative — price levels, distances, regime classification, VIX thresholds — is pre-computed in Python and injected into the prompt as hardcoded facts. The AI structures the narrative; it does not calculate or infer market data.
This architecture is the primary defense against hallucination. The AI receives statements like "ES futures are 141 points above the Put Wall" rather than raw numbers to compute from.
What GEXLOG Is Not
- Not real-time. All data is batch-generated pre-market and post-market. No live intraday recalculation.
- Not institutional GEX. True dealer positioning requires data not available from standard options chains. GEXLOG uses GEX-weighted OI as a proxy — directionally valid, not institutionally precise.
- Not multi-expiration. Nearest-expiration focus means levels may differ from aggregate providers. This is intentional, not a gap.
- Not trade signals. GO/CAUTION/WAIT guidance reflects regime context, not entry/exit signals. All strategies presented use defined risk.