In May 2026, Milliman published research calculating three statistical measures of predictive power for Classic FICO, VantageScore 4.0 (Vantage), and FICO Score 10T (FICO 10T) using data from one of the three major credit bureaus. The statistical measures were the same statistics proposed for use by Freddie Mac and Fannie Mae—government-sponsored enterprises (GSEs)—in the initial solicitation for modernized credit scores: the Kolmogorov-Smirnov statistic (KS-statistic), the bottom decile lift, and the Gini coefficient.
At the time of the analysis publication, Freddie Mac and Fannie Mae had already released data with their single-family loan performance datasets, allowing the mortgage industry to evaluate the performance of Classic FICO and Vantage. However, data for FICO 10T were not publicly available. On July 1, 2026, Freddie Mac and Fannie Mae released data for FICO 10T. The datasets provided include loan-level data and performance for mortgages scored by Classic FICO, VantageScore 4.0, and FICO 10T.
Milliman analyzed this data in order to evaluate the relative fit of the modernized credit scores FICO 10T and Vantage, replicating the prior study with an independent dataset of 46 million mortgages from origination years 2013 to 2025. This paper provides a summary of the results of the analysis.
This research was commissioned by FICO.