Private beta By invitation
Equity factor risk,
from the inside out.
An advanced proprietary factor model over US, European, Japanese and global universes — with ex-ante risk decomposition, currency overlays, constrained portfolio construction and reproducible notebooks. Built for professionals who need to know where the risk is, not just how much.
The model
Transparent by construction.
Value, size, momentum, quality, cash generation, country and currency factors, estimated daily on point-in-time universes reconstructed from fund memberships — survivorship-aware, with immutable data snapshots.
- Universes
- Russell 1000 / 3000, S&P 500/400/600, MSCI World, MSCI Japan, MSCI Europe, S&P Europe 350 — point-in-time membership, monthly snapshots.
- Estimation
- Fama-MacBeth daily cross-sections on local-currency returns; factor and specific risk with shrinkage-estimated covariance.
- Currency
- Local-currency estimation with explicit FX overlays — risk reported in USD or EUR with currency as a decomposed component.
- Discipline
- Coverage policies with named breaches, outlier screens, ETF cross-checks on benchmarks, hash-manifested data snapshots.
Capabilities
One platform, four questions.
What am I exposed to, where does my variance come from, what should I hold, and can I re-derive any number myself.
Exposure & risk
Factor exposures, variance decomposition with per-factor contributions, currency overlays, benchmark-relative analytics.
Optimizer
Min-variance and max-diversification with weight and factor-exposure caps — solved over the full universe in seconds.
Benchmarks
Constituent-aggregated benchmark returns on point-in-time memberships, cross-checked against ETF total returns.
Notebooks
Reproducible marimo notebooks on shared company workspaces — every published number can be recomputed.
Positioning
A risk framework — deliberately.
Fabrisk Equity is not an alpha discovery or backtesting platform. Factor returns and risk premia are descriptive statistics of the model, exposure history is point-in-time from launch, and every data-quality guarantee is stated rather than implied. The methodology note is available to participants on request and inside the platform.