We do not predict the market.
We build observable structures.

Introduction: Market Entropy and the Abyss of Information Asymmetry

1.1 The thermodynamic nature of financial markets and the law of mortality

The most fundamental challenge confronting modern financial engineering is the persistent increase in disorder—entropy—displayed by the market as a vast complex system. According to the second law of thermodynamics, the entropy of an isolated system increases over time and ultimately leads to the system’s “heat death.” The stock market is no exception to this physical law.¹ Countless companies are listed—created—but over a long time horizon most are eliminated, delisted, or stagnate as their returns converge toward the market average. In that sense, listed companies also face a mortal destiny.

Market entropy manifests as uncertainty of information and volatility. Price movement resembles Brownian motion close to a random walk, which is one of the principal obstacles to prediction.² When an external shock such as an economic crisis or pandemic arrives, market entropy changes abruptly and can render established linear economic models ineffective.³ In such a high-entropy environment, finding an orderly signal capable of producing alpha—excess return—is as difficult as finding a needle in a field of sand.

Liquidity also behaves in a way analogous to gravity. Capital tends to concentrate where returns are high, repeatedly producing a local process of self-organization that lowers entropy, followed by a collapse process that raises it again.⁴ The core of successful investment is therefore to capture the low-entropy intervals in which temporary order forms inside the broader disorder.

1.2 The structural disadvantage of the individual investor

Within this disorderly universe, individual investors are placed at a structural disadvantage. Institutional investors and market makers convert information asymmetry into an advantage by using high-performance algorithms and immense datasets.⁵ Individual investors, in contrast, are exposed to unrefined news, rumor, and biased information—that is, to noise.

Information-asymmetry theory explains that an information gap creates adverse selection: participants with superior information continue to earn excess returns, while those with inferior information bear the loss.⁶ The simple technical analysis and intuitive trading on which many individuals rely no longer provide an adequate defense in a modern market dominated by complex derivatives and high-frequency trading. Individuals are like a raft swept into an immense wave, drifting without knowing where an anchor can be set.

1.3 The origin of VERASTRUM: proposing Financial Cosmology

The Financial Cosmology Engine was born from this background and proposes a change in the fundamental perspective—a paradigm shift—in how the stock market is viewed. We define the market not as a mere venue for transactions, but as an immense universe governed by creation and extinction, gravity and orbit. By applying principles from complex-system physics and astrophysics to financial data, the service seeks to discover the order hidden inside market disorder and translate it into language an individual investor can understand.

VERASTRUM is not the familiar kind of service that predicts which stocks will rise. It is a navigational instrument that uses a cosmological lifecycle model to investigate the phase in which each security currently resides and the location of the SuperCluster in which the gravity of powerful capital flow has formed. Our aim is not to name a stock that is “certain to rise,” but to provide individual investors with a data- and logic-verified reliable base on which they can stand and trust their own judgment.⁵

Establishing a New Worldview: A Cosmological Interpretation of Markets

2.1 Price and volume as spacetime: a financial application of general relativity

In the VERASTRUM worldview, price and volume correspond to the spacetime and matter of the physical universe. Einstein’s general theory of relativity explains that matter bends spacetime and curved spacetime tells matter how to move. The corresponding market principle is that capital—volume—distorts the chart—spacetime—and the distorted chart—trend—directs how price moves.⁷

Traditional technical analysis interpreted price movement only as a line on a two-dimensional plane. VERASTRUM adds the concept of density through volume and interprets the market as a three-dimensional structure.

Gravitational lensing: a security receiving an immense inflow of capital draws in surrounding liquidity and forms a powerful upward trend. Like a massive celestial body bending light, this appears as market attention concentrating in one location.

Event horizon: a large buying wall at a particular price can prevent price from falling below that level. We compare this boundary to the event horizon of a black hole. If the boundary is crossed—if support breaks—price may fall rapidly, like a descent toward a singularity.⁸

2.2 The lifecycle of a security: from Nebula to BlackHall

Just as stars are born from interstellar matter, evolve through stable and unstable stages, and end through explosive release or collapse, securities also have distinct lifecycles. VERASTRUM uses AI modeling to classify every security into a cosmological phase. The phase names below use the current VERASTRUM nomenclature; the market-state descriptions and investment postures retain the source manuscript’s meaning.

This lifecycle classification allows an investor to recognize whether a prospective position is a growing star or an aging star approaching instability, and thereby manage risk at its source.¹

2.3 Gravitational-field theory: the mass of capital and the distortion of capital flow

Just as Newton’s law of gravity explains the attraction between two bodies through mass and distance, buying force in the stock market can be explained through the scale of capital—mass—and the separation between the current price and the average acquisition price—distance.

VERASTRUM’s AI calculates and maps this invisible gravitational field across the listed universe. The purpose is not merely to inspect the visible shape of a chart, but to understand the dynamics of the forces operating behind it.

SuperCluster M3: Structural Stability and the Event Horizon of Committed Capital

3.1 The astronomical characteristics of Messier 3 and their financial metaphor

SuperCluster M3, a core grade in the VERASTRUM service, was inspired by the real astronomical object Messier 3, a globular cluster. M3 contains roughly half a million stars packed into a spherical structure. Despite an age of about 11.4 billion years, it remains structurally stable and contains numerous variable stars that serve as standards for measuring cosmic distance.¹²

In financial markets, a SuperCluster M3 security shares the following characteristics of the M3 cluster:

• High density of capital: just as stars crowd the center of a globular cluster, institutional and committed capital is densely interwoven in an M3 security, creating substantial downward rigidity—a reliable base that does not collapse easily.¹⁴
• Structural stability: M3 has existed since the early formation of the galaxy. Likewise, an M3-grade security is not simply a theme-driven spike; it follows a stable orbit grounded in fundamentals or powerful long-term momentum.
• Standard candle: just as the variable stars of M3 are used to estimate cosmic distance, an M3-grade security acts as a benchmark for measuring the strength of the market’s leading sector.¹⁵
• Blue straggler: M3 contains stars that appear young and blue even inside an old cluster. A stellar collision or transfer of mass rejuvenates them.¹⁵ In finance, they symbolize securities that turn around and release powerful price energy through restructuring, a merger or acquisition, or the development of new technology. VERASTRUM is designed to discover these blue stragglers.

3.2 Tracking the whale’s average price and the physics of support

The most important technical logic in calculating an M3 grade is the reverse estimation of the whale’s average price. From a market-microstructure perspective, price is not merely the sum of buys and sells; it is determined by the imbalance between aggressive market orders and passive limit orders.¹⁰

VERASTRUM combines order-flow analysis and volume distribution to understand where committed capital is likely concentrated.¹⁶

Institutional footprint: large tick-level transactions are analyzed to distinguish retail panic buying from algorithmic institutional accumulation. An institutional accumulation zone becomes a powerful gravitational center.

Position leveling: the current price is quantified according to its location relative to the estimated average price of committed capital.

• Level A — M3 Zone: immediately above the estimated average price. This is the safest region and the place where upward energy is concentrated.
• Level B — Momentum Zone: the region in which price begins to erupt; the domain of momentum entry.
• Level C — Overbought Zone: the region in which price has moved far from the estimated average and profit-taking supply can emerge.

An M3 grade means that price remains in Level A while committed participants defend their cost basis by building a powerful buying wall comparable to an event horizon.¹⁷

3.3 The leveling system: grading position through topology

This leveling system goes beyond a simple price comparison and includes topological analysis. We model a price chart not as a two-dimensional time series but as a three-dimensional manifold. In that representation, an M3 grade corresponds to an inflection point immediately before curvature is maximized in the positive direction.

M3 therefore does not mean “a stock that has already risen a great deal.” It means “a stock that has completed its preparation to rise but has not yet been fully exposed to the public.” It is an optimal balance that offers the individual investor both the necessary margin of safety and upside potential.

Probability Density Functions and Five-Day Spacetime: Architecture of the AI Prediction Engine

4.1 The end of determinism and the probabilistic advantage: applying GMM

VERASTRUM does not present a specific target price for an individual security. A deterministic prediction such as “Samsung Electronics will trade at KRW 70,000 tomorrow” is not only impossible in a financial market governed by quantum-like uncertainty; it is deceptive.¹⁸ We focus instead on probability.

Our AI engine learns from several months of historical data and generates a probability density function—PDF for future stock returns. Financial time-series data does not follow a normal distribution. It exhibits a non-normal, fat-tailed distribution in which several local patterns coexist. Gaussian Mixture Models—GMM are used to model this structure.¹⁸

GMM decomposes a complex market-return distribution into the sum of k component Gaussian distributions:

Through this model, the AI separately calculates the probability that the present market condition belongs to an upward distribution, a sideways distribution, or a crash distribution.¹⁹

4.2 The probability of a 5% rise in five days: thresholds of short-term volatility and trend

Why five days, and why 5%? This setting is an optimal spacetime selected with both statistical significance and trading efficiency in mind.

Five-day spacetime—one trading week: five trading days are the minimum interval required to complete one weekly candle. This horizon filters the noise of very short-term scalping while excluding some of the macro uncertainty of long-term investment. It also aligns with short option-expiry cycles and is therefore well suited to reflecting supply-and-demand effects from the derivatives market.¹⁹

A 5% rise—the threshold of momentum: 5% is not merely a return objective; it is a threshold at which a volatility breakout is confirmed. A high probability of rising at least 5% within five days suggests that the security has left an ordinary random walk and entered a strong momentum interval.²¹

Our model does not predict that “the price will be X won after five days.” It calculates a probabilistic edge in the form: “the probability density that the price will be at least 5% above its current level after five days is 70% or greater.”

4.3 Combinatorics of indicators and optimization of multidimensional vector operations

A single technical indicator—such as RSI or MACD—cannot defeat the market by itself. The reasons are multicollinearity between indicators and overlapping signals.²² To overcome these limitations, VERASTRUM calculates combinations of indicators.

The AI combines hundreds of technical indicators, capital-flow variables, and macroeconomic indicators to generate trillions of possible signal combinations. It then uses genetic algorithms and deep learning to identify the “super-indicator combinations” with the highest win rate.²³

For example, rather than a simple golden cross, the following high-dimensional logic is executed:

These compound conditions reveal subtle correlations that human intuition cannot capture. By allowing each indicator to compensate for the blind spots of the others, they sharply improve predictive accuracy.²⁵ This is the essence of VERASTRUM’s high-dimensional logic.

Top-k Ranking Algorithms and the Refinement of Information

5.1 Differentiable Top-k classification and loss-function design

The output of AI modeling is a distribution of upward probabilities across thousands of securities. To transform this enormous dataset into information an investor can act on, a Top-k algorithm is applied. The list is not produced by simply cutting the universe at the highest raw probabilities. Instead, the model uses differentiable Top-k classification, a modern machine-learning technique that optimizes the performance of the highest-ranked k-member group from the training stage onward.²⁶

This means the model learns to maximize the expected return and robustness of the Top-k portfolio as a whole rather than the point accuracy of each individual security. The selected securities are consequently constructed with lower correlations to one another, so the portfolio can remain stable even when some individual forecasts fail.

5.2 Preventing overfitting through purging and embargo in cross-validation

The greatest enemies of financial-data modeling are overfitting and look-ahead bias. A model that fits only historical data perfectly is useless in live trading. To prevent this, VERASTRUM uses strict purging and embargo cross-validation techniques.²⁸

Purging: breaks temporal dependencies between training and test data to prevent data leakage.

Embargo: excludes data for a defined interval immediately after the test data from training, preventing the model from learning future information in advance.

Only models that survive this strict validation process are used to calculate five-day upward probability. A VERASTRUM Top-k report therefore does not merely recite past glory; it is the result of testing future possibilities as conservatively and objectively as possible.

5.3 Translating quantitative results into qualitative reports

The final computed data is organized into a Top-k graded report that a user can understand. The report removes unnecessary mathematical complexity from the viewing layer and provides intuitive visual material.

• SuperCluster M3 universe map: visualizes, as a map of stellar clusters, the sectors into which market capital is currently concentrating.
• Upward-probability gauge: displays the probability of a 5% rise after five days in the form of a speedometer.
• Capital-flow leveling chart: provides a topological map of the current price’s location relative to the estimated average price of committed capital.

The report does not compel the reader to buy a security. It communicates the facts—for example, that a security has entered a SuperCluster M3 orbit and has a high upward probability—so the user can make the final judgment.

A New Paradigm for Individual Investors: Building a Reliable Base

6.1 Probabilistic grounds for reaction, not prediction

Countless stock-picking channels and market experts claim to sell “accurate prediction,” but the market is intrinsically unpredictable. VERASTRUM does not sell prediction. We provide the grounds for a response.

The principal reason individual investors fail is the absence of a standard they can trust. A small decline triggers fearful selling; a small rise provokes a greed-driven chase. VERASTRUM supplies the mathematical and logical foundation needed to quiet that psychological instability. A user can verify that “the security I own is SuperCluster M3, is trading above the estimated average price of committed capital, and carries an 80% five-day upward probability,” allowing ordinary short-term noise to be endured and the underlying signal to be followed.

6.2 Democratizing information and establishing independent investment through study

Our objective is to eliminate information asymmetry. By delivering to individual investors the high-dimensional quantitative analysis and capital-flow data once available only to institutions, we seek to level the tilted playing field.⁶

VERASTRUM does not, however, want a user to remain a passive beneficiary. By refusing to state a guaranteed future return, we encourage the user to undertake a measure of independent study. The AI completes 90% of the analysis—security discovery, capital-flow analysis, and probability calculation—while the user fills in the remaining 10% by reviewing financial statements, searching material news, and applying a personal investment philosophy.

Through this process, the user grows from a mere consumer of information into a smart investor. VERASTRUM is not a cane on which to lean blindly; it is a reliable base that helps the investor stand and walk independently. This is the genuine reduction of information asymmetry we pursue, and the highest value we can provide to an individual investor.⁵

6.3 Conclusion: the financial future presented by VERASTRUM

VERASTRUM is a scientific answer offered to individual investors frustrated by the limitations of technical analysis and the barrier of information asymmetry.

The insight to read the stock market through a cosmological lifecycle, the powerful capital-flow metaphor of SuperCluster M3, and the refined probability engine formed by Gaussian mixture modeling and Top-k algorithms together open a new world of investment analysis.

The age of investment governed only by feeling and rumor is ending. In VERASTRUM’s Financial Cosmology—governed by probability, physics, and data—the investor becomes an explorer able to navigate the immense universe of the market. From the reliable base we provide, the investor can survey a wider horizon of opportunity.