October 27, 2026

More Rigorous Validation for Predictive Models

We've refined how FactDecode trains and validates predictive models. Factors are reselected within each historical validation period, with stricter boundaries to prevent future target outcomes from l…

October 22, 2026

FactDecode: Evaluate Factor Selection with Model Performance and Run Comparisons

With so many factors to choose from, which ones should you keep? FactDecode now brings factor screening, Nested OOS model performance, and score-band analysis together in Results. Compare saved Ru…

October 20, 2026

Choose Up to 5 or 10 Factors, with Clearer Results

FactDecode now lets you choose whether your final model can include up to 5 or 10 factors. We've also improved how Results handles cases where the model cannot identify valid factor contributions, ma…

October 27, 2026

Build on Past Analyses with FactDecode

FactDecode now lets you bookmark completed Runs, making important past analyses easier to find. You can also reuse a saved Run's candidate factors and Raw Data selections when preparing your next ana…