And all scores perform the same general function: to use a consumer's credit history to try to predict whether he or she will pay debts in the future.

FICO scores summarize this prediction in a three-digit number that ranges from 300 to 850. People with lower scores are considered higher risks in that they might not repay their debts; people with high scores are considered low risk.

For decades, FICO kept a relatively low profile. Then, in 1995, Fannie Mae and Freddie Mac, the government-sponsored giants that form the backbone of the mortgage industry, decided to start using the FICO score in their computer programs to help decide which consumers qualify for mortgages bought and sold by the companies.

"That's really when it took off," says Evan Hendricks, the founder of PrivacyTimes, who has testified before Congress on credit-scoring matters.

Next came the housing boom. As many banks loosened their lending requirements, they drastically cut back on their own internal underwriting departments, relying more and more on the FICO score to determine whether a consumer was worthy of credit. Soon, Fair Isaac was marketing its score to companies as varied as health insurers and casinos, advertising it as a low-cost way for companies in an array of markets to decide who got credit and who didn't.

That is how a company that most people have never heard of, and even fewer understand, came to play such a pivotal role in the day-to-day functioning of the modern economy.

"FICO is the wizard behind the curtain of the economy," Matt Fellowes, a scholar at the Brookings Institution, told BusinessWeek in 2008. (Fellowes did not respond to calls and emails seeking comment.)

The Basics II: What's in that special sauce?

To arrive at a consumer's credit score, FICO considers a number of different factors, including how many different kinds of credit accounts a consumer has, how much available credit the consumer tends to use and whether the consumer makes payments on time.

FICO updates its scoring model every two or three years, much like Microsoft periodically releases new versions of software, Quinn says. It makes the updates partly to take advantage of new kinds of data and to react to changes in consumer behavior.

For example, many consumers drastically reduced their credit card debt after the financial crash of 2008. In the face of such a broad behavioral change, the model must change too, since what may have been a low credit utilization rate in 2006 might have been simply average by 2009.

"Over time, degradation can occur to any scoring model," says Clifton O'Neal, a spokesman for TransUnion, one of the three national credit bureaus.

The newest version of the scoring model is more precise because it considers more types of consumers. For example, some people have little in the way of credit histories. Comparing those people to all consumers as a whole would leave them with low credit scores, even though many people with little credit history make perfectly fine credit risks. The same goes for people with unpaid bills in their past. In some cases, that may mean three unpaid mortgage payments, which could mean the person is a bad credit risk now. Another person's unpaid bills might be for small medical expenses that she was unaware of.

By dividing people with similar credit histories into different groups and comparing them against people in those groups, FICO is able to give more-accurate predictions of consumers' willingness and ability to pay.

"If you were to build just one scorecard on the whole population, people with previous delinquency would all plummet to the lowest part of the score range, because compared to most other people, they look really risky," says Quinn. "But if you build a scorecard based just on delinquencies, someone with lots of delinquent payments now is riskier than someone with one late payment six years ago. So it helps you find people with less risk."

Whereas the previous model divided consumers into 10 categories, the new one uses 16. That should help lenders make more fine-tuned decisions, according to FICO.

"With FICO 8 score, FICO scientists were allowed for the first time to break the blueprint of our original FICO scoring algorithm," Jason Sprenger, a FICO spokesman, told via email. That creates "a segmentation that rendered significantly more predictive credit scores than was possible before."

Compared with previous versions, FICO 8 goes easier on people who have missed payments on small debts under $100, which often include medical debts. The Commonwealth Fund estimates that 14 million Americans struggle with medical debts that they believe are erroneous, The Wall Street Journal reported, and the Federal Reserve estimates that half of all debts in collection arise from medical bills.

That means the new FICO scoring model could help millions of consumers improve their credit scores and reduce the amount of money they pay in interest.

"One of the good things in 8 is that debts under $100 are not going to have the same impact," says Hendricks, of PrivacyTimes. "I think it's safe to say a majority of those are medical debts."

On the other hand, FICO 8 considers people with high credit utilization rates to be somewhat higher credit risks than previous versions. That may hurt the credit scores of people who use relatively high levels of their available credit. "So if you carry high balances, you're probably going to lose a little more points than before," Quinn says.

The two tendencies may cancel each other out, Quinn says. Banks probably won't become any more loose with credit under the new model, and they probably won't lend to more people. Instead, lenders will subtly shift their definitions of which consumers qualify for which products.

Reading tea leaves

Most people haven't noticed any difference at all from FICO 8, and it's not because their credit histories are impervious to the changes. Rather, major lenders have been unusually slow to switch to the new scoring model.

In fact, the model was originally called FICO 08 because the company had planned to implement it in 2008. (Because of a lawsuit FICO brought against the Equifax credit bureau, the release was delayed until early 2009.) Now that it's going on four years since the original release date, FICO simply dropped the zero. Three years after the actual release, most major lenders still have yet to adopt it. And FICO 8 is entirely absent in the mortgage market.

But good luck getting confirmation of that fact from FICO, the credit bureaus or any major lenders, which use consumer data to make the scores but largely decline to tell the public what they do with it.

"There's no publicly available data on it," Hendricks says. "But that's my sense of it, that it has been slower."

Figuring out exactly what's happening with FICO 8 is an exercise akin to reading tea leaves. Try calling FICO to ask how implementation of FICO 8 is going, and they say you should call the credit bureaus. Call the bureaus, and they refer you to the banks. Call the banks, and they refuse to talk, saying their scoring models are proprietary.

The upshot: This is an industry that relies on consumers' data to function and that dictates the finances of millions of people. And yet it arguably operates with almost zero accountability to consumers.

"It's not as transparent as it needs to be," Hendricks says. "All this information should be filed publicly so we know what's going on. But we don't."