0.899
Forecast error vs. EWMA (1.000)
Evidence, not hype.
Performance
Built on Discipline.
Explore validated forecast evidence, changing market regimes and risk context in a clearer, easier-to-read format.
Explore validated forecast evidence, changing market regimes and risk context in a clearer, easier-to-read format.
EigenPoint publishes forecast-accuracy evidence from disciplined, pre-registered testing rather than promotional return figures. The figures below measure how accurately the engine forecasts volatility, not investment returns.
Live market display is provided by a third-party market-data widget and may be delayed depending on exchange and symbol.
Forecast error vs. EWMA (1.000)
Assets improved over HAR
Assets where our core design added accuracy
Live forecasts logged since Oct 2026
Forecast error is measured relative to EWMA (1.000; lower is better); the industry-standard HAR model scored 0.918. Figures come from a pre-registered holdout test: 2023–2026 data on nine assets (eight exchange-traded funds and Bitcoin) was set aside, never used in development, and evaluated once under pass/fail rules fixed in advance. They measure volatility-forecast accuracy, not trading or investment performance. Full methodology is available to qualified counterparties under NDA.
A disciplined process and rigorous historical evaluation give qualified institutions more to examine than promotional claims alone.
