# Options Dealer Positioning — Overview

> Personal SpotGamma/QuantVue-style options-positioning pipeline. Pulls institutional-grade options chain data (CME + OPRA via DataBento, ETF chains + spot via Schwab), computes per-strike dealer-positioning metrics (GEX, walls, gamma flip, vanna, charm, max pain, L1-L4), renders a multi-symbol multi-DTE web dashboard, and tracks SpotGamma's own published levels daily for empirical comparison.

## Why this exists
Tradeable dealer-positioning signals — call walls, put walls, gamma flip, max pain — are usually paywalled (SpotGamma $99–$299/mo, Polygon $199/mo, QuantVue more for live). With my own DataBento + Schwab subscriptions I already pay for the underlying data. This project turns that into the same dashboards retail vendors charge for, plus a daily comparison framework against SpotGamma's published levels.

## Status (2026-05-26 EOD)
**Live, working, daily reproducible.** Multiple iterations of fixes applied through the day; current build at `dashboard_visual_20260526_142345.html`.

### Coverage (9 symbols)
| Symbol | Source | Cost |
|---|---|---|
| **ES** | DataBento CME `ES.OPT` + 9 weekly parents (`E1A-E1D`, `EW1-EW4`, `EW.OPT`) | $0 |
| **NQ** | DataBento CME `NQ.OPT` + 8 weekly parents | $0 |
| **GC** | DataBento CME `OG.OPT` (gold smart symbol) | $0 |
| **RUT** | DataBento OPRA `RUT.OPT` + `RUTW.OPT` (cash index, direct match to SG) | ~$7/mo |
| **SPX** | DataBento OPRA `SPX.OPT` + `SPXW.OPT` | ~$7/mo |
| **NDX** | DataBento OPRA `NDX.OPT` + `NDXP.OPT` | ~$7/mo |
| **IWM** | Schwab REST `/chains` (Russell ETF — kept in parallel to validate RUT/IWM scale factor) | $0 |
| **SPY** | Schwab REST `/chains` (S&P 500 ETF) | $0 |
| **QQQ** | Schwab REST `/chains` (Nasdaq ETF) | $0 |

**All spot prices** come from Schwab `/marketdata/v1/quotes` (free).

**Monthly cost:** ~$22 in DataBento OPRA usage; CME free in subscription tier; Schwab free. (Was briefly down to $7 when only SPX OPRA was enabled; bumped to $22 when user accepted RUT+NDX cost for direct SpotGamma comparison.)

### Major features as of EOD 2026-05-26
- **SpotGamma indicator-string parser** (`sg_levels.py`) — 14-field positional decode of their daily morning string
- **3 view modes** in dashboard: **Ours** (default), **SG** (SpotGamma's published levels), **Compare** (side-by-side with delta)
- **Daily SG-vs-ours tracking** persisted to `reports/level_tracking.jsonl` for empirical scoring
- **L1-L4 metric** by OI ranking (reverse-engineered to match SG's L-level methodology for ETFs)
- **Abs Gamma Strike** (gamma-weighted magnet level, complement to L1)
- **Hover tooltips** (ⓘ icons) on every metric — ELI5 explanations inline
- **1% wall proximity badge** — fires when spot enters the SpotGamma 1% rule zone
- **Implied 1-day range card** (ATM IV × spot × √(1/252) — 78% of sessions close inside)
- **4-up bar charts** per strike: Net Delta, Net Gamma, Net Vanna, Net Charm
- **Per-expiration max pain** + γ% contribution (anticipates gamma-decay regime shifts)
- **Zero-GEX zones** secondary level list
- **Auto-archive** of older dashboard HTML files into `reports/archive/`

## Chapters
1. [[01 - Architecture & Data Sources]] — how Schwab, DataBento, WSL, and OneDrive fit together
2. [[02 - Concepts (ELI5)]] — what gamma / GEX / walls / max pain actually mean
3. [[03 - Math & Sign Conventions]] — formulas, multipliers, sign conventions, IV recovery
4. [[04 - Symbology & Gotchas]] — `OG.OPT`, weekly parents, RTY-not-in-CME, RUTW/SPXW/NDXP weeklies, etc.
5. [[05 - Dashboards & Run Guide]] — `dashboard_visual.py` workflow + view tabs + SG daily input
6. [[06 - Cost & Budget]] — DataBento usage-based vs flat-rate alternatives
7. [[07 - Future Work]] — intraday-OI corrected GEX, C-level math, settlement-spot replay
8. [[08 - SpotGamma Playlist Findings]] — full SpotGamma methodology survey from 99-video learning series
9. [[09 - SpotGamma Tracking & Reverse Engineering]] — daily comparison framework + L/C level math reverse-engineering status
10. [[10 - AWS S3 Hosting Setup]] — step-by-step one-time AWS account setup for auto-publishing the dashboard to a shareable URL (manual, you do everything in console)
11. [[11 - AWS S3 Setup via Claude (delegated)]] — alternative path where you create a temp admin IAM user, Claude runs the AWS CLI setup commands, then you delete the temp user (~5 min total)

## Related notes
- [[Schwab Options Data API]] — original 2026-05-11 plan note (now superseded; Schwab plays the spot+ETF-chain role only)
- [[ATS 2.0 Trading System]] — QuantVue's commercial system this approximates

## Code
- Live: `/home/dlitt/projects-personal/schwab-options/` (WSL ext4, fast)
- Backup: `/mnt/c/Users/dlitt/OneDrive/claude/projects-personal/schwab-options/` (OneDrive periodic copy)
- Venv: `~/.venvs/schwab-options/` (Python 3.12, httpx + databento)
- Credentials (NEVER in code, NEVER on OneDrive):
  - `~/.schwab/credentials.json` — Schwab client_id + secret + tokens
  - `~/.databento/credentials.json` — DataBento api_key
- Daily SG input: `~/.schwab/sg-levels-YYYY-MM-DD.txt` — paste SG's "Copy All" string each morning