3
120+
4
Methodology tracks
From statistical foundations to applied return decomposition
Three structured tracks take you from the core mathematics of latent variable methods through dynamic factor models and into applied return decomposition and risk attribution — at the depth institutional work actually demands.
The Learning Path
A sequential curriculum that builds from statistical theory through to full LVRM
implementation, each stage grounded in the questions practitioners actually face when
working with real asset return data.
%!d(template.templateFloat=01)
Foundations
Statistical Foundations
4
Core modules
Before latent variable methods make sense, the probabilistic scaffolding that supports them must be solid. This stage covers maximum likelihood estimation, the geometry of covariance structures, and the distinction between observed and unobserved components in a return series. Participants who arrive with gaps in this foundation consistently report that addressing them here changes how they read the rest of the curriculum. The emphasis is on building intuition alongside formal facility, so that later estimation choices feel motivated rather than arbitrary.
Three to four weeks
%!d(template.templateFloat=02)
Factor Models
Factor Model Architecture
6
Core modules
Static and dynamic factor models are introduced here as the structural backbone of most latent variable approaches to return data. We examine how factor solutions are identified, why rotation indeterminacy is a genuine interpretive challenge rather than a technicality, and how the number of latent factors is selected in practice using information criteria and scree-based diagnostics. The comparison between principal component approaches and confirmatory factor structures is treated in detail, because practitioners routinely encounter both and the differences matter for how results are interpreted and communicated to non-specialist stakeholders.
Four to six weeks
%!d(template.templateFloat=03)
Estimation
Latent Variable Estimation Methods
8
Core modules
Kalman filtering, state-space representation, and Bayesian approaches to unobserved component estimation are the technical core of this stage. Each method is introduced with its assumptions stated explicitly and its failure modes examined honestly. Participants work through the specification of measurement equations and transition equations, and learn how to diagnose model misspecification from residual diagnostics and likelihood surfaces. Regime-switching frameworks for identifying structural breaks in latent factor behaviour are introduced toward the end of this stage as a natural extension of the state-space approach.
Five to seven weeks
%!d(template.templateFloat=04)
Application
Applied Portfolio Construction
7
Core modules
With estimation methods established, this stage turns to the applied questions that motivated the curriculum: how do decomposed return components inform risk attribution, how are latent volatility structures used in position sizing and drawdown analysis, and how do practitioners communicate factor-based risk decompositions to investment committees and compliance functions that may not share the same quantitative background. Case studies drawn from multi-asset return data are used throughout, and participants are expected to work through decomposition exercises using their own analytical tools.
Four to five weeks
%!d(template.templateFloat=05)
Capstone
Research Capstone
1
Reviewed capstone
The final stage asks participants to apply the full methodology to a return dataset
of their choosing, producing a structured research document that identifies latent
factors, justifies the estimation approach selected, reports attribution results
with appropriate uncertainty quantification, and discusses the limitations of the
analysis honestly. Capstone submissions are reviewed by the platform's
quantitative research team, who provide written feedback grounded in the same
standards applied to peer-reviewed empirical work. Participants who complete this
stage leave with a documented analytical output they can reference in institutional
settings.
Three to four weeks
Choose Your Methodology Track
Contact
for track details
Designed for analysts who need to build or consolidate the statistical groundwork before engaging with advanced latent variable estimation — ideal for those transitioning into quantitative roles or returning to formal methods after time in practice.
- Introduction to covariance structure analysis
- Static factor model specification and identification
Probabilistic foundations of return modelling
Maximum likelihood estimation principles
Rotation indeterminacy and its practical consequences
Diagnostic tools for model evaluation
Recommended
Practitioner Track
Contact
for track details
Built for working quantitative analysts and institutional researchers who already
have a grounding in statistical methods and want structured engagement with dynamic
factor models, state-space estimation, and applied risk attribution techniques.
Kalman filter specification and implementation
Bayesian approaches to unobserved component models
Regime-switching frameworks for latent factor dynamics
Research Track
Contact
for track details
Intended for research-oriented professionals at asset management firms, central banks, or academic institutions who require deep engagement with frontier latent variable methods and ongoing access to the platform's research community and methodology updates.
Full curriculum access across all methodology stages
Priority access to new methodology releases
Dedicated feedback from quantitative research team
Track and Curriculum Questions
The right track depends on where your analytical foundations currently sit — and where your institutional work requires you to go
Whether you are consolidating statistical groundwork, engaging with state-space estimation for
the first time, or conducting frontier research into high-dimensional latent factor structures,
there is a track designed for the specific depth your work demands. Reach out to discuss which
starting point makes sense for your background and objectives, or explore the methodology
underpinning the curriculum before committing to a track.
Rigorous methods grounded in current research
Every technique introduced across the three tracks is held to the standard of current empirical literature, with honest treatment of assumptions, failure modes, and the limits of each estimation approach when applied to real asset return data.
Peer-level engagement across institutional backgrounds
Participants include analysts from asset management firms, central banks, and research departments — the diversity of institutional perspectives sharpens how latent variable methods are interpreted and applied across different analytical contexts.
Applied outputs that connect directly to institutional work
The capstone process produces a documented analytical output grounded in the participant's own return data, reviewed by the platform's quantitative team to the same standard applied to peer-reviewed empirical research submissions.