A number that reflects how likely it is that a person will pay back a loan. A credit score is usually calculated from a credit scorecard and is engineered to range along from some arbitrary scale (such as between 200 and 800). A bank or other financial institution picks a cut-off score. Loan applicants with scores less than the cut-off are rejected, and loan applicants with scores greater or equal to the cut-off are accepted. A bank may refine the strategy by offering lower interest rates for applicants with higher scores.
A correctly engineered scorecard can predict both the likelihood that a person with that score will repay a loan and the number of people who will receive that score. For the bank, there is a trade-off: a higher cut-off score means fewer default loans but few total loans. The bank must determine the value of each good (repaid) loan, the cost of each bad (default) loan, and calculate the percentage of bad loans it is willing to accept.
The common argument against credit scoring is that it prevents a loan officer from using judgment and intuition in making loan decisions. The people making these arguments assume that such decisions would be made in favor of someone who would be otherwise rejected by a scorecard. What is often overlooked is that a loan officer might just as easily reject an applicant who might be accepted by a scorecard. With a scorecard, a bank can't turn down an applicant simply because the applicant doesn't "look" like a good applicant to the loan officer. The decision algorithm of a scorecard, unlike a loan officer, is open to scrutiny.