How a credit score works — a number strangers use to guess whether you'll pay them back

Personal Money 3 min 80 sources

A credit score isn't a report card on your character. It's a prediction, built from your past borrowing, that lenders buy to decide your rate and your limit — and five things move it.

Key takeaways

A credit score is a prediction a lender buys to guess whether you'll repay — not a report card on your character.
Five things move it: payment history (the biggest), how much of your limit you use, how long you've borrowed, your mix of credit, and how recently you've applied.
Checking your own score never hurts it, negative marks fall off after about seven years, and the formula itself is a company's product — not a fixed law about you.

You apply for a loan, a card, or an apartment, and somewhere a three-digit number decides how you’re treated. Most people never learn what it is or how it moves. Here’s the machinery.

A credit score is a number, usually between 300 and 850, that predicts how likely you are to pay back a loan on time [7][43]. It is not a verdict on whether you’re responsible, honest, or good with money in general. It answers one narrow question a lender is asking: if we hand this person credit, will they repay it? [43]

The number comes from a credit report — a file that companies called credit reporting companies (or credit bureaus) build about your borrowing [63]. They collect the loans you’ve taken, the credit limits you hold, your balances, whether you paid on time, any accounts sent to collections, and public records like bankruptcies [63]. A credit score is then a number calculated from that report to summarise the risk [43].

The most widely used brand of score is FICO, made by the Fair Isaac Corporation [16][9]. There are several versions, and each weighs your report slightly differently, so the exact number can vary depending on which one a lender pulls [43][11]. But they all read the same five things about you.

Payment history — whether you’ve paid past bills on time — is the single most important factor [7]. Late or missed payments can lower your score sharply [7]. Credit utilization comes next: how much of your available credit you’re actually using [7]. Using more than about 30% of your limit tends to drag the score down [7]. Then length of credit history (how long you’ve held accounts), credit mix (whether you handle both instalment loans like a mortgage and revolving credit like a card), and new credit (how recently and often you’ve applied) [7].

That last one explains a common confusion. When you apply for credit, the lender makes a hard inquiry — a request to see your report to judge the application — and because scoring models watch how recently and how often you apply, hard inquiries can nudge your score down [12]. But checking your own score is a soft inquiry, and soft inquiries never affect it [12]. Looking at your own number costs you nothing [12].

The ranges sort roughly like this. On the 300–850 scale, FICO treats a score below 580 as poor, and 670 to 739 as good [16][7]. The average US score sits around 740 [7]. Lenders don’t just approve or reject on this number — they price you with it, setting your interest rate and your credit limit by where you land [43][18]. The Consumer Financial Protection Bureau groups borrowers into tiers, from deep subprime (below 580) up through prime (660–719) and super-prime (720 and above) [31].

Two things about the timeline matter. Negative information — a late payment, a default — can stay on your report for up to seven years [24]. And you generally can’t have accurate negative information removed; anyone promising to erase current, accurate, negative marks is likely running a scam [27]. Positive history, by contrast, can stay longer and keeps working for you even after an account is paid off and closed [24].

The score isn’t fixed law, either. A high-profile fight has been running over which scoring model mortgage lenders must use — a reminder that the formula behind the number is a business product set by companies and regulators, not a natural fact about you [33]. Different models, different numbers, from the same underlying file [11].

The practical edge: the score is downstream of behaviour a report captures. Pay on time, keep balances well under your limits, and don’t open a rush of new accounts at once — and check your own report at least once a year for errors, which you have a legal right to do [62][30]. What you can’t do is game a number that’s just a summary of a real track record.

02 · Lesson · why it matters

The number that decides for you before you're in the room

A credit score is a stranger's prediction about you, dressed up as a fact about you — and the arrangement that builds it was chosen, not found.

A judgment you’re not present for

You walk into a bank, or fill in a form for an apartment, and a decision is already half-made. Before anyone hears your reasons, a three-digit number has spoken. It set your rate, your limit, sometimes whether the door opens at all.

The strange thing is how absent you are from your own case. You didn’t submit the evidence. You didn’t choose the criteria. A file was built about you by companies you never hired, read by a formula you’ve never seen, and the result met you at the counter fully formed. This is worth sitting with, because it’s the shape of a lot of modern life: you are constantly being scored by systems you’re not in the room for.

What the number actually is

Here’s the first thing to see clearly. A credit score is not a measure of your worth, your honesty, or your discipline. It answers one narrow question a lender asks: will this person pay us back? It’s a prediction, built from a record of what you did before.

That distinction matters because we feel it as a verdict. A low score lands like a moral grade — you failed. But the machine isn’t grading your character. It’s guessing your future from your past, the way a weather forecast guesses tomorrow from a pattern of yesterdays. Mistaking a prediction for a judgment is how people start believing the number is the truth about them, rather than a bet someone placed on them.

Your past self is funding your future self

Now the connection. The score binds two versions of you across time. The person who paid a bill late three years ago is quietly setting the price of the loan you want today. The person paying carefully this month is buying the future self a lower rate. Nothing about you feels connected in the moment — a single late payment feels like an isolated slip. The system treats none of it as isolated.

And it binds you to strangers. The whole apparatus — the bureau that collects the file, the lender that buys it, the formula that scores it — exists for one reason. Lending to people you can’t vouch for needs a shortcut for trust. The score is that shortcut. It lets a bank in another city lend to you without knowing you, which is also what lets you borrow at all. The same machine that judges you is the reason credit is available to a stranger like you in the first place.

The formula was chosen, and it poses as natural

Here is the part that’s easy to miss, because the number arrives looking like a measurement — objective, neutral, simply true. It isn’t. Someone built the formula. A company decides that paying on time matters most, that using more than a slice of your credit counts against you, that a rush of applications reads as risk. Those aren’t laws of nature. They’re choices, encoded, that then present themselves as plain fact.

You can watch this seam open up in public. There’s an ongoing fight over which scoring model mortgage lenders should be required to use — because a different formula gives a different number from the exact same file. If the number were a fact about you, it wouldn’t change when the company changes. It changes because it was always a product: a rule someone wrote, sold, and can rewrite.

That arrangement serves its makers — it’s a business — and it can genuinely help the people under it too. A score that travels lets an honest borrower prove reliability without a personal introduction. Both things are true. The point isn’t to name a villain. It’s to stop treating a chosen rule as the weather.

No single seat sees the whole

Notice how little any one player actually sees. The lender sees a number, not you. You see the number, not the formula. The bureau sees your file, not the loan you’re trying to get or the reason you need it. Each seat holds one slice, acts on it, and the slices never assemble into a full picture in anyone’s hands.

This is why the humble move is to work with the machine rather than against it or in fear of it. You can’t argue with a formula at the counter. But the number is downstream of things a record captures — paying on time, leaving room under your limits, not applying in a rush. And it’s built from a file you have a legal right to see and correct. What you can’t do is talk your character into it. The system doesn’t read character. It reads a track record, and it reads it whether you’re paying attention or not.

So the number that decides for you isn’t a truth about you and isn’t an accident either. It’s a bet, placed by strangers, running on a formula someone chose, built from a past you’re still tied to. You’re a node in it, scored before you speak. The most useful thing is to see the whole shape of it, and hold your own number a little more loosely than the counter wants you to.

03 · Lab · your turn

What the Number Sees

Feed a credit score's five factors and watch the strangers' prediction move, seeing the total as only ever the sum of its parts.

04 · Hope · carry this

The number is only a summary of what you did, which means the pen is still in your hand — every on-time payment is a quiet vote for the version of you that comes next.