Quark develops proprietary risk and regime detection signals built on novel research in mathematical physics. Our models identify crash risk, regime transitions, and structural market shifts days before conventional indicators.
Quantifies the real-time probability of a market regime transition into crisis. Our model detected the onset of the COVID-19 crash 16 trading days before the drawdown began, with 76.5% precision on drawdowns exceeding 5%.
Separates genuine price momentum from microstructure noise and liquidity artifacts. Delivers a real-time signal quality score that tells you whether a move is driven by informed flow or temporary friction.
Monitors the structural stability of asset relationships in real time. Instead of trading on a fixed schedule, our system identifies exactly when portfolio factor exposures have shifted enough to warrant action — reducing unnecessary turnover by 40–55%.
Classifies the current market environment into one of four distinct regimes on a continuous coordinate system. Each regime maps to a measurably optimal allocation strategy, replacing subjective regime labels with quantitative precision.
Combines topological analysis of market microstructure with regime transition modeling. Detects when the geometry of return distributions is shifting before it becomes visible in price or volatility.
Analyzes how market dynamics change across timescales — from intraday through weekly. Acceleration in cross-scale divergence leads conventional regime detection by 2 to 5 days, providing an actionable early warning window.
Multi-source data across equities, derivatives, macroeconomic indicators, digital assets, and foreign exchange. Tick-level through weekly frequency.
Proprietary decomposition isolates signal from noise at multiple timescales and across asset classes.
Seven independent engines analyze regime stability, crash risk, factor structure, and cross-asset dynamics.
Orthogonal signals are combined into a single risk/opportunity surface. Each signal is independently validated.
Daily reports, API feeds, crash probability alerts, and regime change notifications.
Primary crash signal tested across SPY, QQQ, IWM, DIA, and EEM over 2019–2023. 76.5% precision on drawdowns exceeding 5%. Provided 16 trading days of advance warning before the March 2020 crash.
Multi-asset portfolio running live since February 2026. Active positions across equities, digital assets, foreign exchange, and derivatives. Signal-guided allocation and automated risk management operating in real time.
Every signal has a pre-registered hypothesis with defined null conditions, required sample sizes, and multiple-testing correction. Over 55,000 experiments logged. Our validation framework enforces statistical discipline automatically.
The foundational whitepaper. Develops a drift-diffusion framework for decomposing asset dynamics, regime transition modeling via potential barrier analysis, and multi-scale signal construction with connections to 85+ works in the literature.
A self-contained pedagogical treatment bridging mathematical physics and quantitative finance. Designed for practitioners who want to understand the analytical foundations without a physics background.
Survey of 142 papers across 18 thematic areas spanning portfolio optimization, factor investing, regime detection, and cross-disciplinary quantitative methods. Maps the full research landscape informing our signal development.
Describes the architecture of our autonomous portfolio management engine. Signal integration, adaptive position sizing, risk controls, execution logic, and the feedback loop between signals and allocation decisions.
Payoffs linked to the ratio of genuine momentum versus market noise in a given asset or index. Allows passive investors to hedge signal degradation and active managers to isolate and trade alpha-quality regimes directly.
Contracts that pay out when factor structure destabilizes beyond a threshold. Directly prices the cost of diversification failure — the unpriced risk that correlations suddenly reorganize during stress.
Structured notes whose coupon adjusts based on the rate of change in asset relationship structure. Low coupon when factor exposures are stable; rising coupon compensates holders as structural instability increases.
A swap contract whose floating leg references our proprietary crash probability surface across multiple time horizons. Unlike static volatility indices, the reference rate responds directly to changes in regime transition dynamics.
Options with knock-out barriers that respond to market noise intensity rather than price level. Natural risk control: positions self-deleverage when the market becomes untradeable, not when an arbitrary price level is breached.
Recent work on topological data analysis applied to market regime classification. Persistence landscapes provide stable, vectorizable features from return point clouds — directly relevant to our structural breakdown detection capabilities.
Extends classical stochastic volatility with barrier-crossing corrections for tail events. Our crash probability framework builds on this class of methods, adding multi-scale calibration and empirical validation against five major indices.
Our factor stability signals detected structural breakdown in equity correlations 2–3 days ahead of the April 2025 tariff announcements. Defensive rebalancing reduced portfolio drawdown relative to benchmarks that held static allocations.
Internal validation showing that acceleration in parameter divergence across timescales consistently leads conventional regime indicators by 2–5 trading days. Tested on COVID-19 crash, 2022 rate shock, and 2023 banking stress.
Survey connecting Nelson-type stochastic decompositions to signal extraction in quantitative finance. Provides theoretical grounding for separating genuine momentum from market microstructure noise in real time.
Preliminary results from our live deployment show that factor-stability-triggered rebalancing reduced unnecessary portfolio turnover by 45% compared to fixed weekly schedules, with no measurable sacrifice in risk-adjusted returns.
Download a sample report, or reach out to discuss how Quark fits your workflow.
contact@quarkquant.com