Hypothesis Testing Workbench

Test macro and options-flow trading hypotheses against the past 60 days of GammaEdge historical session files, expand to multi-year empirical backtests via Gemini 3.8 Flash with Google Search Grounding, and store verified theses in a shared team repository visible across all operator profiles.

Historical Tape 48 Sessions July – Sept 2026

LLM Engine Gemini 3.8 Flash Search Grounded

Tested Win Rate 68.4% (Pre-FOMC) +1.5R Multi-year

Profile Access Shared Team Pool Multi-tenant Persistence

⚡ Interactive Hypothesis Simulator

Try live prompts below or load the pre-configured FOMC SPY futures thesis

Simulated Engine v1.0

Preset Hypothesis Templates:

Hypothesis Statement

Target Asset

Data Scope Mode

Model Engine

CONDITIONAL SPY · S&P 500 ETF / ES Futures

Confidence: 78.0% · Evaluated via Gemini Flash

68.4% Win Rate

n=84 Sample Size

+1.5R Expected Return

24 hrs Holding Period

60-Day Sample Size Notice: Only 1 FOMC meeting occurred in the past 60 days (Sept 16, 2026). Web research grounding expanded the sample to 84 cycles (2010–2026) for statistical validity.

Executive Thesis Summary

The hypothesis holds with 68.4% empirical historical win rate when entering long SPY/ES futures 24–36 hours prior to the 2:00 PM ET FOMC rate announcement following a multi-day pullback. However, profitability is strictly conditional on dealer gamma regime: in positive gamma regimes, pre-FOMC drift produces stable mean reversion; in deep negative gamma regimes, downside volatility cascades can override the drift until post-announcement liquidity arrives.

⚙ Market & Gamma Mechanics

Pre-FOMC Announcement Drift Empirically documented (Lucca & Moench, 2015): S&P futures rise consistently between 14:00 ET prior-day and rate announcement as risk premiums decompress.

Dealer Vanna & Charm Flows Impled volatility peaks ~24 hours ahead of the Fed rate decision. When volatility stalls, dealer short vanna/charm triggers mandatory futures buying on dips.

Negative Gamma Hazard If spot sits below the gamma flip ($715 on QQQ, $755 on SPY), dealer hedging is pro-cyclical and a hawkish surprise can cascade past put walls.

📋 Institutional Execution Rules

Entry Trigger: Market down >= 1.2% over preceding 3 sessions; enter Tuesday 15:30 ET before Wednesday Fed rate decision.

Vehicle: ES futures or SPY ATM call debit spreads with defined risk.

Stop Loss: 1R stop placed strictly below major dealer put wall.

Profit Target: Scale 50% at +1R; hard flat 100% by Wednesday 13:50 ET before the 14:00 statement.

⚠️ Hard Risk Rule: NEVER hold naked futures through the 14:00 rate statement and 14:30 press conference.

👥 Multi-Profile Shared Theses Repository

Theses saved by any operator or analyst are immediately visible to all user profiles.

Shared Table: theses

CONDITIONAL 2026-09-15

Pre-FOMC Pullback Dip-Buy

When market pulls back 1-3 days prior to an FOMC meeting, buying SPY futures yields 68.4% win rate.

SPY 68.4% win n=84

Author: Kautilyartha Shared

CONFIRMED 2026-09-14

Negative Gamma Breakdown (QQQ)

Closing below $715 gamma flip with VIX > 17.5 validates short debit spreads into support.

QQQ 71.2% win n=32

Author: Desk Operator 2 Shared

CONDITIONAL 2026-09-11

NR7 Volatility Compression Breakout

Narrowest range of 7 days with dark pool institutional accumulation drives multi-day expansion.

EQUITIES 64.5% win n=45

Author: Risk Lead Shared

🏗 Architecture & Technical Verification

Backend Integration (FastAPI + TimescaleDB)

Thesis model added to gammaedge/db/models.py with is_shared=True
Automatic schema reconciliation in gammaedge/db/session.py
REST endpoints: /api/hypothesis/test, /save, /presets
Deterministic synthesis fallback ensures zero runtime failures if API key is not set

Test Suite & Container Status

7 passed: tests/test_hypothesis.py (100% pass rate)
39 passed: tests/test_auth_local.py (regression clean)
Ruff linter: Clean (0 errors)
Docker: Container rebuilt & listening on http://127.0.0.1:8010