Macro Architecture

The Total Approach: Systems vs. Time Series

A fundamental error in quantitative analysis is treating an asset as a mere "time series"—a sequential list of past prices. A time series is just a one-dimensional shadow cast by a much larger, unseen object. The Total Approach fundamentally shifts the paradigm: we do not forecast isolated time series; we model the market as a massive, interconnected Dynamical System.

The Time Series Fallacy

Traditional time series analysis assumes an asset's future can be predicted strictly from its own history. It relies heavily on localized lags and self-variance. Because it is completely blind to the underlying structural mechanics, a purely time-series-driven model fails catastrophically during regime shifts, black swans, or external liquidity shocks.

The Systemic Paradigm

A system is a dynamic network of interconnected state variables. An asset is merely a single node reacting to structural forces, latent macroeconomic factors, and mathematical lead-lag relationships with other nodes. By analyzing the entire system's topology and state space, we execute based on the "engine's" mechanics, not just its exhaust.

The Mathematical Foundation

This philosophical shift necessitates a profound change in the underlying mathematics. Instead of using localized autoregression to predict the next sequential step in a vacuum, a robust backend trading engine models the stationary relationships across the ecosystem:

Isolated Time Series $y_t = \alpha + \phi y_{t-1} + \epsilon_t$
System State-Space $Y_t = \beta X_t + \Gamma Z_t + \epsilon_t$

In the systemic equation, the vector $X_t$ represents latent principal components (macro forces) and $Z_t$ represents the mathematically detected market regime. The residual $\epsilon_t$ now represents a true structural arbitrage opportunity rather than unexplainable random noise. When processing asynchronous logic across thousands of portfolios simultaneously, relying on these systemic state variables is the only way to mathematically guarantee the architecture does not over-leverage on fragile, highly correlated anomalies.

Dashboard Navigation: Use the sidebar to explore the theoretical math in the Foundational Lenses (Section II), and then interact with the live execution simulations in the Execution Strategies (Section III).
Theoretical Lens 01

Pairwise Relations: The "Leash" Effect

While standard correlation measures short-term, directionless volatility mimicking, Cointegration represents a long-term, mean-reverting economic equilibrium. Imagine a person walking a dog on a leash; they may wander apart temporarily, but they are ultimately tethered together.

Cointegration Visualization

Theoretical Lens 02

Latent Factors: Core Drivers

Principal Component Analysis (PCA) allows us to isolate the "invisible forces" (such as macroeconomic shifts or interest rate decisions) that drive the broader market. This lens shatters the illusion of diversification.

Total Market Variance Distribution

More than 70% of equity portfolio movement is explained by the First Principal Component.

Theoretical Lens 03

Topological Lens: Mapping Market Keystones

Network theory enables us to map asset centralities. If a Keystone Asset—a highly central node with massive institutional interconnectedness—experiences a shock, the contagion spreads aggressively, threatening the ecosystem's structural integrity.

Keystones Clusters Peripherals
Theoretical Lens 04

System Dynamics: Predicting Fragility

When all assets suddenly begin moving "in lockstep," the system has become extremely fragile. The Absorption Ratio quantifies this fragility, serving as an early warning system before actual price drops materialize.

Systemic Fragility vs. Index Returns

Execution Insight 01

True Diversification via Orthogonal Factors

Constructing a portfolio where weights are distributed across independent mathematical risk factors (PCA), not just different company names. This ensures survival during targeted macroeconomic shocks.

Sim: Naive vs. True Diversification

Execution Insight 02

Statistical Arbitrage Execution

Monitoring the Z-score of the spread between cointegrated assets. If the Z-score exceeds thresholds, the automated engine shorts the overvalued asset and buys the undervalued one, profiting upon mean-reversion.

Live Trading Algorithm

PnL: $0
Execution Insight 03

Lead-Lag Information Flow

Monitoring primary assets. If a major breakout occurs in a highly liquid market, immediately execute directional trades in secondary lagging markets before market makers update quotes.

Sim: Propagating Alpha

Execution Insight 04

Contagion Hedging Protocols

Dynamically monitoring the volatility of the top 5% most central nodes to trigger portfolio-wide risk-off protocols days before a localized failure cascades into a systemic contagion.

Sim: Cascading Failures

Execution Insight 05

Capital Rotation Tracking

Using rolling volume-weighted correlations to track the net flow of liquidity between macro sectors, front-running the shift from growth to safe-haven assets.

Sim: Economic Phase Shifts

Execution Insight 06

Markov Regime Detection

Employing Meta-strategies. When the scanner detects a transition to a "High Volatility" regime, the system automatically disables trend-following algorithms and activates mean-reversion engines.

Sim: Adaptive Algo Yield

Background colors denote the hidden market regime. Adaptive Algo switches logic during volatile chops (Red).

Execution Insight 07

Systemic Fragility Gauge

When the Absorption Ratio crosses historical bounds, systematic engines aggressively reduce gross exposure and acquire portfolio insurance weeks before the actual drawdown occurs.

Sim: The Crash Precursor

Execution Insight 08

Macro Inefficiency Discovery

Event-driven statistical arbitrage. The millisecond new data arrives, algorithms short assets that are computationally "behind" in pricing the shock and hedge with assets that have already priced it in perfectly.

Sim: Shock Pricing Velocity