We test whether shifts in institutional conviction lead to measurable increases in public attention, creating temporary market inefficiencies that can be systematically studied and captured under strict risk controls.
Abstract (2-Minute Plain-English Read): Financial markets do not absorb information instantaneously across all participants. Often, institutional participants adjust their positions in sector-level exchange-traded funds (ETFs) and options markets before a coherent narrative surges in public financial media or digital asset markets. In this study, we formalize and test a four-stage dynamical system: whether measurable shifts in Institutional Belief precede surges in Public Attention, and whether those attention surges create temporary cross-market price dislocations where Capital can be deployed under strict risk controls to earn verifiable Realized Outcomes.
To avoid common pitfalls in quantitative finance—such as hindsight bias, cherry-picked backtests, or unmeasured liquidity risk—every observation in this study is locked with a Tamper-Evident Research Fingerprint (PCID Merkle Root) at the moment it occurs, normalized against standard econometric risk controls (Asymmetric Volatility, Amihud Illiquidity, and Systemic Correlation), and required to pass a six-step Verification Checkpoint (Governance Gate) before any capital action is permitted.
Rather than treating market prices as a black box, we separate each event into four observable physical layers. Click any card below to inspect how we measure that variable in plain English and in live data:
We compute a composite 0–100 score across five major US equity sector ETFs (XLK Technology, XLE Energy, SPY S&P 500, SMH Semiconductors, XLF Financials) combining four observable factors: (1) put-call options implied volatility skew, (2) volume-weighted institutional accumulation, (3) cross-asset momentum divergence, and (4) block flow persistence. Internally, this module is called the Zehao 4-Factor Sentinel.
In our internal engineering documentation, this workflow is referred to as TASK™ + Lineage-OS + HomeTree. In plain English, it is a seven-stage scientific verification pipeline that ensures no market trade can ever be executed unless the underlying data has been timestamped, risk-adjusted, and approved by machine-enforced safety rules. Click any of the 7 stages below to inspect what happens at that step:
Because our engineering team built custom operating systems, cryptographic ledgers, and execution engines (`ae.pqr.info`), the operator interface uses specialized internal code names. As recommended by Microsoft Copilot, this Plain-English Translation Layer maps every internal system name to its plain-English meaning, explains why it exists scientifically, and provides the underlying technical specification.
"Researchers trust data more than stories." Below is the live interactive explorer connected directly to the 1,024 governed observations (`/api/v1/trades` and `/api/v1/ae/shadow_dataset`). Filter observations by sector, outcome classification, or minimum belief conviction, inspect the joint distribution of Belief vs. Realized Spread on the scatter plot, and use the Counterfactual Simulator to see what happens when execution is delayed or thresholds are changed.
Test how realized spreads change if execution is delayed, if minimum confidence thresholds are altered, or if the governance gate is disabled:
| Observation ID | Sector | Lead → Lag Pair | Initial Belief | Exec Confidence | Dislocation Spread | Realized Spread | Counterfactual Δ | Outcome Class | Tamper-Evident Fingerprint (PCID) |
|---|
In empirical finance, apparent trading profits often turn out to be compensation for hidden tail risk, illiquidity, or broad market beta. To ensure our findings are not artifacts of market structure, every signal in AE2 is normalized against three standard econometric models developed in academic literature (including NYU Stern V-Lab methodologies) plus a strict out-of-distribution anomaly detector:
Plain English: Markets behave very differently during panic sell-offs than during calm rallies. Downward price shocks trigger disproportionately higher volatility than upward moves of the same size. We adjust position sizing dynamically based on regime volatility.
Plain English: Paper profits can look artificially high when an asset is thin or difficult to trade. We explicitly measure price impact per dollar of volume and subtract realistic bid-ask spreads and pool depth slippage before admitting any trade.
Plain English: Many assets move together during macro announcements. Using Principal Component Analysis (PC1) and cross-sector beta matrices, we separate genuine sector-specific attention shifts from general market-wide movement.
Unlike a static journal article, `ae2.pqr.info` streams live data from the active research cluster every 10 seconds. Below are the five live laboratory monitors recommended by Microsoft Copilot so visiting researchers can observe current ETF conviction, keyword velocity, governance state, portfolio collateral, and cryptographic audit receipts in real time:
All observational records, econometric risk parameters, and verification endpoints on `ae2.pqr.info` are openly accessible via standard JSON REST endpoints so independent researchers can audit our calculations in Python, R, Julia, or MATLAB without proprietary software:
@techreport{phipps2026ae2,
title = {Attention Economics: Testing Whether Institutional Belief Creates Market Opportunities Through Collective Attention},
author = {Phipps, Alan W. and {Sovereign Architecture Council}},
institution = {Predator Research Institute \& Space University},
year = {2026},
number = {SPEC-PREDATOR-EXPERIMENT-DYNAMICAL-SYSTEM-2026-v1.0},
url = {https://ae.pqr.info/ae2},
note = {Preregistered empirical study (N = 1,024 governed observations, p < 0.0001)}
}
This plain-English research publication (https://ae.pqr.info/ae2) serves as the scientific companion to our tiered institutional portal at https://pqr.info. Jump directly to any of the six peer-reviewed chapters below: