The Science Behind VantageScore

Learn more about VantageScore model innovations

Criteria for conventional scoring models (many still in use today) may date back more than 30 years. They reflect the data and methods of THAT time, not today. VantageScore is the modern credit scoring model for the digital age and beyond.

The incorporation of machine learning technology and the insights gained from analyzing credit behavior over time, via the use of trended credit data, make it possible for VantageScore 4.0 to score more people with greater accuracy.

Machine Learning Transforms Credit Scoring

VantageScore 4.0 is the first and only generic, tri-bureau credit scoring model to apply the latest in computer-driven, machine-learning technology. Specifically, it leverages this advanced technique to generate a more predictive credit score among consumers who have had no credit activity in the past six months, but have relevant information in their credit file that is more than six months old.

Data scientists at VantageScore leveraged machine learning techniques to obtain an even deeper look into the credit histories of these consumers, who are often overlooked by lenders, and examined these specific details to identify additional patterns that corresponded with positive financial behavior or increased risk.

By leveraging machine learning, VantageScore scientists were able to develop 50,000 attributes for examining sparse credit histories resulting in more than a 10 percent performance lift amongst those with Dormant credit histories and more than a 30 percent performance lift amongst those with “No Trades,” as compared with the performance generated by conventional credit score models.

Attributes discovered in this process also meet certain standards necessary for inclusion into a generic credit scoring model, such as an attribute’s ability to be translated into reason codes.

Trended Credit Data Insights

It has long been a Catch 22 of credit: consumers who do not regularly use credit typically struggle to get approved for new credit accounts.

For years, consumers who didn’t have a recent track record of credit cards, loans or other types of borrowing were either unable to get a loan at all, forced to pay higher interest rates or required to turn to other alternatives, such as pawn shops or payday lenders.

VantageScore 4.0, the new standard for modern credit scoring, is the first to use trended credit data to analyze credit behavior over time, providing deeper insights into borrowing and payment patterns. Together with the computational power of machine learning, our model is able to score approximately 37 million 1 more consumers than other commercially available models, without sacrificing predictability.

Predictive performance has been significantly improved, especially in the Prime and Super Prime segments. For Prime originations, trended data can provide a 20 percent improvement in predictive performance when used in combination with traditional static attributes.

With VantageScore 4.0, lenders gain access to nearly 37 million potential customers who would not be able to obtain a credit score with the older scoring models still used by many financial institutions, including all mortgage lenders.

  • Based on a 2018 study performed on a nationally representative random sample of 15 million consumers.

Research and Whitepapers

Access the latest industry reports and studies, covering everything from emerging market trends to technology innovations.

VantageScore Market Adoption Study

A recent study conducted by global management consulting and research firm Oliver Wyman reported that usage of VantageScore credit scores is deep and mainstream and occurs in all credit industries except the mortgage market. The study found that 12.3 billion VantageScore credit scores were used in a 12-month period between July 2018 and June 2019, a 20 percent increase in usage over the prior year. Since June of 2015, VantageScore usage has grown approximately 20% per year.

Read the Study

Improved Assessment of Credit Health Using Trended Credit Data

This white paper introduces trended credit data and explains why the analysis of consumer credit behavior over time is a stronger indicator of credit risk than a snapshot of consumer credit behavior.

Read the Study

2020 Model Performance Assessment of VantageScore 4.0

To aid in transparency, VantageScore performs and publicly publishes annual model assessments. This 2020 report represents the third annual model performance assessment ofVantageScore 4.0 (which was launched in April 2017). The model is the first and only tri-bureau credit scoring model to incorporate, for superior performance, both trended credit data and leverage machine learning.

Read the Study

Browse More Research

From regulations to research, VantageScore Solutions presents a closer look at the latest issues facing the credit industry.

  • Based on a 2018 study performed on a nationally representative random sample of 15 million consumers.

contact us vantagescore icon

Get In Touch with Us

VantageScore.com is loaded with educational materials about credit and scoring and information on our models and how to use them. However, if you need additional help, or if you’re a lender looking to speak with someone directly, please get in touch with us here

  • Based on a <a href="https://vantagescore.com/education/blog/conventionally-unscoreable-population-grows-to-40-million-consumers">2018 study</a> performed on a nationally representative random sample of 15 million consumers.