FICO just released a “comparison” of VantageScore 4.0 vs. FICO Score 10T—and it is a jaw-dropping masterclass in statistical misdirection.
Rather than comparing the two models head-to-head on the same dataset, the authors play a deceptive game of inference:
- They compare FICO 10T to FICO Classic on one dataset…
- Then compare VantageScore 4.0 to FICO Classic on a different dataset…
- And somehow conclude how VantageScore 4.0 stacks up against FICO 10T.
That’s not a model comparison. That is a logic leap - and a misleading one.
Even worse: their analysis relies on a cherry-picked one-year window (April 2016 - April 2017) instead of the full 10-year dataset published by the GSEs. No explanation given. But the intent is clear: make FICO 10T look uniquely strong.
This is not robust or sound analysis. It is bad science - and it undermines trust in fair, transparent credit model evaluation.
The truth? At least three independent, published analyses - by JPMorgan Chase, Bank of America, and KBRA - found that VantageScore 4.0 outperforms FICO Classic.
Want the facts? Download the full 10-year dataset from Fannie Mae and Freddie Mac and see for yourself. Real performance speaks louder than selective storytelling.
Authored by Dr. Rikard Bandebo, EVP, Chief Strategy Officer and Chief Economist
