Four Signals Before the Headlines

Every tool investors use today explains what already happened. Factor models decompose past returns. Technical indicators derive output from prior price action. News alerts arrive after the market has already moved.

But markets move before news breaks. Information enters market microstructure — through order flow, volume anomalies, and price behavior — before the event becomes public. Two decades of peer-reviewed research from Georgia Tech’s Scheller College of Business, published in the Review of Financial Studies and the Journal of Financial Economics, have documented that these pre-disclosure behavioral anomalies are real, detectable, and carry measurable significance. (For a full treatment of the academic research and how it converges with Bimini’s independent findings, see our white paper: Detecting What Markets Know Before You Do.)

Bimini’s Intelligence Refinery was built to detect them in real time — using only public data. When detections meet statistical thresholds, the system surfaces them as Advance Indicator Notices, delivered with a Chain of Logic that shows the full reasoning trail.

These four case studies demonstrate the system in action across different data domains: market microstructure, government economic releases, regulatory filings, and cross-market correlation.

---

Case Study 1 — Actual Detection

Raytheon and Ukraine: The Market Knew Two Months Early

In late 2023, Bimini detected a Ghost Pattern in Raytheon Technologies (RTX) — a persistent two-month accumulation anomaly in the stock’s microstructure. No public catalyst existed. No news, no earnings surprise, no analyst upgrade. The pattern’s duration and structure were consistent with sustained informed buying by participants with foreknowledge of a material event.

The Pentagon subsequently announced it had observed Russian troop accumulation on the Ukrainian border and believed an attack was imminent. Raytheon — a primary manufacturer of Patriot and Iron Dome missile defense systems — jumped significantly on the announcement.

Bimini’s system had identified the signal two months before the public catalyst.

What this demonstrates: Cross-domain correlation at work. A behavioral anomaly in a defense equity — detected through market microstructure alone — was subsequently confirmed by a government announcement. The microstructure signal and the government signal pointed to the same event, from structurally independent data domains. All inputs were publicly observable.

---

Case Study 2 — Illustrative Example

The Data That Disappears: EIA Oil Revision Delta

Every Wednesday at 10:30 AM Eastern, the Energy Information Administration (EIA) publishes its Weekly Petroleum Status Report ^[1] — crude oil inventories at Cushing, Oklahoma, well head counts, refinery utilization. Markets react within seconds. Oil futures, energy Exchange-Traded Funds (ETFs), and related equities reprice immediately.

Three days later, the EIA quietly revises Wednesday’s numbers. The original figures disappear from the EIA website, replaced by the revised data.

The original version is not preserved in any major financial data terminal or intelligence platform we have evaluated.

The delta between what the government reported, how the market reacted, what the government later corrected, and how the market responded to the correction — that data exists nowhere. Until now.

Bimini’s Intelligence Refinery captures every government data release at the moment of publication as an immutable, timestamped record. When the revision arrives, the system computes the delta and correlates both versions against market behavior.

This pattern repeats across every major government economic release: Bureau of Labor Statistics (BLS) Nonfarm Payroll revisions, Bureau of Economic Analysis (BEA) Gross Domestic Product (GDP) advance-to-final revisions, Consumer Price Index (CPI) seasonal adjustment corrections.

What this demonstrates: Proprietary data creation from public sources. The raw data is public, but the act of storing the original alongside the revision creates a dataset that cannot be replicated retroactively. A competitor starting today would need to wait years to build what Bimini already has. This is a compounding data asset — every week adds another data point, every revision adds another base rate observation. The dataset grows automatically.

For Registered Investment Advisors (RIAs), the compliance value is immediate: advisors can document client decisions in the context of data that was later revised, protecting against hindsight bias in compliance reviews. The Chain of Logic surfaces the original report, the revision, the delta, the market reaction to each, and the historical base rate for similar revision events — a compliance-grade record.

---

Case Study 3 — Actual Detection

Tesla: The Ghost Pattern That Preceded Musk’s Billion-Dollar Buy

In early September 2025, the Intelligence Refinery detected a Ghost Pattern in Tesla Inc. (TSLA) — a behavioral anomaly in the stock’s microstructure consistent with unusual accumulation from an unknown source. The detection was derived entirely from publicly available market data. The stock began rising sharply on September 10.

On September 12, Elon Musk executed 25 separate purchase transactions totaling 2,568,732 shares of Tesla stock — approximately $1 billion. Three days later, on September 15, a Form 4 was filed with the Securities and Exchange Commission (SEC) disclosing the trades ^[2]. Tesla continued to rise sharply through October 1.

A Form 4 is a legally required SEC disclosure. Under securities law, corporate “insiders” — defined as officers, directors, or shareholders holding more than 10% of a company’s stock — must report any transaction in their company’s securities within two business days. As Tesla’s CEO and a greater-than-10% shareholder, Musk’s purchases were routine, lawful transactions disclosed through the standard regulatory process. There is no allegation of wrongdoing. The intelligence value is in the timing: the microstructure signaled accumulation before the filing made the trades public.

Bimini’s Ghost Pattern detection had identified the anomalous microstructure behavior before the insider trades occurred — and well before the Form 4 filing made them public. Most advisors learned about the trades days later, after the move was already underway.

What this demonstrates: Ghost Pattern detection identified anomalous accumulation in Tesla’s microstructure before the insider’s trades and before the regulatory filing made them public. The Intelligence Refinery continuously monitors market microstructure across 104+ securities and correlates detected anomalies with subsequent regulatory filings — Form 4 insider transactions, 13-F institutional holdings, 8-K material event disclosures — to confirm what the microstructure signaled. The Ghost Pattern was the leading indicator; the Form 4 was the confirmation.

---

Case Study 4 — Actual Detection

GOOG–NQ: The Market Told Us Twice

In July 2024, the system detected simultaneous Ghost Patterns across two structurally independent instruments — Alphabet Inc. (GOOG) on the equity side and NASDAQ 100 futures (NQ) on the derivatives side.

On July 23, 2024, a Ghost Pattern was identified on GOOG with a 13-day duration — a persistent, large-scale selling anomaly consistent with informed positioning ahead of a material event. Concurrently, the system detected an intraday Ghost Pattern on NQ futures reflecting massive selling pressure. On August 1, the signal intensified: pre-market intraday patterns revealed panic-like selling and hedging activity across both NASDAQ 100 and Russell 2000 futures.

The catalyst was a federal antitrust ruling against Google — not yet public when the patterns were first detected. The NASDAQ 100 subsequently declined 12%.

This case was published and timestamped on X (formerly Twitter) before the resolution of the catalyst — part of 28+ publicly validated detections in Bimini’s auditable track record.

What this demonstrates: Cross-market validation — the strongest form of signal confirmation. GOOG is a major component of the NASDAQ 100. When a Ghost Pattern appears on an individual equity that constitutes significant index weight, and a concurrent Ghost Pattern appears on that index’s futures contract, the system interprets this as layered detection from structurally independent instruments. The equity pattern suggests knowledge of a company-specific catalyst; the futures pattern suggests informed participants understand the catalyst is material enough to move the broader index.

This cross-market propagation mechanism maps directly to peer-reviewed research by Dr. Suzanne S. Lee at Georgia Tech’s Scheller College of Business (2012, Review of Financial Studies), which demonstrated that information enters individual stock microstructure and propagates to the index level. Bimini arrived at the same finding independently through live market observation, then discovered the academic validation.

---

The pattern across the patterns

These four case studies span different data domains — market microstructure, government economic releases, regulatory filings, and cross-market derivatives. Each demonstrates a different capability of the Intelligence Refinery. Together, they illustrate the platform’s core architecture: ingesting data from eight distinct domains, transforming each through a proprietary economic lens, and correlating across domains to surface intelligence that exists nowhere else.

Every detection uses only public data. Every assessment is delivered with a Chain of Logic — a compliance-grade reasoning trail showing which data sources contributed, what correlations were drawn, and what historical base rates support the finding. For financial advisors, this is the document they file in the compliance folder. For the Chief Compliance Officer, this is the audit trail that makes adoption possible.

The question these case studies answer is the same question advisors face every morning: which way is the wind blowing? The difference is that now, for the first time, a system can answer it — with evidence, reasoning, and a track record that is public and auditable.

---