/laplace-2/images/laplace_logo_horiz.png){kind=logo}
Financial markets are not just volatile—they are fundamentally non-stationary. Relationships evolve. Correlations break. When an asset, sector, or theme moves into a new regime, yesterday’s signals can lose their relevance.
Many quantitative approaches assume relationships hold over time. That assumption is weakest when conditions change fastest: during regime transitions, structural breaks, and periods of heightened uncertainty.
Take stocks and bonds. They are negatively correlated—until inflation takes hold and the correlation turns positive. Add one dimension, and a simple relationship turns complex. Then the questions multiply. Does it hold by sector or industry? How do currencies, especially the US dollar, reshape it? Does it vary by region? Where do commodities fit?
AI models must adapt as these relationships shift. PSIMON is a specialized AI that is built from the ground up to solve this problem.
A regime is a distinct period of behaviour that shapes volatility and returns, whether for an individual security, a sector or a theme, and our models detect and measure each one as it emerges. Our multidisciplinary team is focused on building specialized algorithms for one of quantitative finance’s hardest problems: identifying regimes in real time.
Our research centres on this core capability: detecting regime changes for each individual asset while capturing the dynamic interdependencies across multiple financial time series. Together, these show how an asset behaves on its own and among its peers, and how long that behaviour persists. Our cutting-edge research has been featured in academic journals and at conferences.
The payoff for investment managers: see which of the assets you follow are changing, understand how they are likely to behave, and focus your research and attention where it matters, when it matters.
Financial market data raises two fundamental challenges: modelling temporal behaviour inherent to financial time series, and accounting for complex interdependencies among the broader market, sectors, and individual assets. Asset Regime Intelligence addresses both with three key capabilities:
Every asset moves through its own regimes. PSIMON® learns them from the asset’s own history, mapping its distinct behavioural states and natural regime structure. Our ensemble models combine supervised and unsupervised machine learning to capture the nonlinear relationships across multiple financial time series.
Not every regime change matters. Each trading day, PSIMON detects when an asset, a sector or the market moves from one regime to another, using cross-correlation analysis and regime-switching probabilistic analyses. It sets aside the shifts unlikely to matter and flags those that may affect the asset’s near-term volatility risk and positioning outlook.
A regime is only meaningful when viewed in context. PSIMON measures each regime by its statistics and persistence, then compares it with the asset’s peers: the sector, the broader market, and user-defined watchlists. The result is two signals: a volatility risk outlook for each asset, plus a positioning signal within each list it belongs to.
PSIMON’s Asset Regime Intelligence leverages ensemble forecasting methods to read every holding and watchlist name in context: against its peer group, its sector, and against the market as a whole.
Under current conditions, PSIMON weighs a range of potential outcomes, drawing on cross-sectional and time-series data across broader macro and financial factors, sectors and individual securities. It then turns that analysis into positioning signals within each peer group or thematic watchlist, so you can judge every name from several angles.
The result: a sharper read on which assets need your attention now.
Our rigorous process emphasizes robustness, adaptability, and practical relevance:
Stress-Tested for Robustness: Models are evaluated across varied market environments, including periods of market stress and dislocation.
Non-Stationary Optimization: Regime structures adapt as asset behaviour and market relationships evolve.
False-Positive Control: Model design and validation aim to reduce noisy signals and momentum-chasing bias.
Continuous Evolution: Models are refined as new evidence and market conditions emerge, with disciplined testing against prior environments.
Know when major changes are forming for an asset you follow, before their full effect is obvious using traditional methods. PSIMON provides a unique input into your research process, alerting when key relationships change, when correlations break down, and when sectors and holdings move into more turbulent (or calmer) phases.
Every signal leads back to your holdings and watchlists. PSIMON flags the assets most likely to see a shift in positioning or risk outlook so your research time can focus on what matters.
Want to see it on assets you follow?
PSIMON’s signals look ahead over different horizons. The volatility risk outlook covers the next five trading days, while the positioning outlook typically holds for weeks to many months. Both rest on the same foundation: each trading day, PSIMON reads the regime, or distinct behavioural state, that each asset is currently in.