Dealing with fraud is a red queen game: The fraudsters can keep trying until they find what gaps are there in your system, and will sometimes communicate with each other: Part of our defense system involved infiltrating some of those spaces and seeing the guides that were being sold to try to commit fraud in our platform. Meanwhile you will still have a false positive rate, and getting it all the way to zero means crazy fraud. Most people just don't get to see how much fraud is stopped before they know it exists. This isn't just for financial institutions: You'd be surprised by how much credential stuffing is attempted at, say, any very large streaming site which charges a subscription.
Without looking in, it's hard for me to say exactly how successful their security team is, but being as big as they are, and having probably thousands of people whose only job is to do fraud on their platform, winning has to be pretty hard.