MaxMind minFraud detects fraud more reliably
A few months of testing and the verdict is in: MaxMind minFraud flags more transactions made with stolen credit cards as high risk than my own fraud scoring code. minFraud does a lot I can’t: Checking the bank BIN location against the billing address, the location of the customer’s phone number against the billing address or if the IP address is a known proxy; if the username or e-mail address are in its database of high risk accounts; if the e-mail has been reported by other users as high risk.
The last one is really amazing — they must have built up quite a large user base considering I’ve gotten many notifications of transactions I ran through minFraud later reported as fraudulent by other users. This weekend I updated my risk scoring code to remove ineffectual checks and weigh the minFraud score as the most important component in the overall score. I now outright deny credit card transactions that are flagged with the highest MinFraud score.
