How Credit Scores Are Built and Why the Same Debt Looks Different

What a score actually measures

A credit score is a summary statistic produced by compressing a file of credit history into a single number between 300 and 850. It is a prediction, produced by a statistical model, estimating how likely a borrower is to repay.

That framing matters. The score is not a measure of whether you are a good person, a good citizen, or financially responsible. It is a narrow model output trained on outcomes from a particular population. It measures repayment behaviour on credit accounts, and it has nothing at all to say about your income, your savings, whether you pay rent on time, or whether you are a good steward of money.

Where the model uses data other than what it was trained on, the result is a genuine fair-lending question, and one that has been litigated. For most people the practical issue is narrower: understanding what moves the number so you can make an accurate report about yourself.

The five factors

Every mainstream model scores these, with weights that differ by model:

Payment history is weighted most heavily, typically around 35%. Whether you pay on time. This is the factor that behaves most as people expect it to.

Amounts owed is usually second, around 30%. This is your total balance relative to your credit limits. Note the word relative: it is a ratio, so owing $5,000 on one card can score better than owing $1,500 spread across three if the first is a small fraction of its limit.

Length of credit history is around 15%. The average age of your accounts. This is the factor that most rewards patience and is least responsive to effort.

Credit mix is around 10%. Having a variety of account types — credit cards, a mortgage, instalment loans — tends to score slightly better than only one type. This is the factor most sensitive to the specific formula, and some models treat it as negligible.

New credit is around 10%. Recently opened accounts and recent hard inquiries.

Two observations worth making.

Weights are ballpark figures and differ by model. Treat published percentages as approximations rather than as an explanation of any particular score.

Your score depends on which bureau’s file is used, and which model the lender runs. The same person can have genuinely different scores at different bureaus on the same day, with no error anywhere.

Why two models disagree

The two dominant models use overlapping but different data and different formulas.

FICO is built around data from the three national credit reporting companies and weights payment history heavily. Variations of the base FICO score exist for mortgages and for credit cards, and each version is computed differently — so “the FICO score” is often an oversimplification of which one is meant.

VantageScore is built on a different data foundation and, importantly, has incorporated rent and utility payment history reported by participating providers — data that does not enter any traditional credit file.

So a person who has never missed a payment but has paid rent on time for years may see a meaningful gap between their VantageScore and FICO. Neither is broken. They are answering slightly different questions from different data.

This matters practically, because lenders do not all use the same model. A score that is comfortably above one threshold can sit below another.

Utilization: the fastest lever

Two distinct ratios are described as “utilization”, and they behave differently.

Statement balance to limit, measured on your statement date, is what most models use. This is the one with a fast, visible effect: paying your balance in full before the statement closes lowers the reported ratio without changing anything else about your accounts.

Revolving utilization, measured more continuously, is used by some models and is harder to influence.

Two practical consequences:

Lowering reported utilization can improve a score within one billing cycle, which makes it the fastest lever available.

Paying your balance in full and also keeping a small balance across limits is the commonly recommended approach — the reasoning is that consistently reporting a small utilization demonstrates you are not carrying balances.

A related point: closing an old account can hurt your score, by reducing total available credit and by shortening your credit history if it was your oldest account. The intuition that unused credit is good credit is wrong.

The order of operations

If a score needs improving, the interventions have very different timelines.

Paying down before the statement closes can move the number within one cycle. This is the only fast lever.

Correcting a reporting error takes as long as the dispute process takes, often 30 to 45 days, and can be worth far more than anything else on the list. Errors are common: payments reported as missed, accounts that are not yours, balances after an account is closed, duplicate accounts. Each of the three bureaus will investigate and must complete an investigation even where you supply documentation.

Deleting a small collection entry may be possible depending on your jurisdiction. In some states and under newer federal rules, if you pay a debt that has already been removed from your file, the furnisher must also request deletion of the negative entry. It requires being in contact with the collection agency and knowing the rule. It is worth several dozen points in some cases.

Changing behaviour to improve payment history takes years, because the model’s longest memory is around seven years and good behaviour accumulates gradually.

Older errors age out on their own schedule.

Frequent misconceptions

Paying off a collection improves your score. The effect depends on which version of the report is being scored, which varies by model. Some versions do not count paid collections at all. Check before assuming.

Closing cards improves your score. It usually lowers it slightly, by reducing total credit and total available credit. Generally leave old, unused accounts open.

Checking your credit hurts your score. Checking your own report is a soft inquiry with no effect. Applying for credit is a hard inquiry.

You need a certain number of accounts. There is no target. Carrying five cards you do not need increases your total exposure and adds nothing.

Credit score myths aside, paying off high-interest debt improves your finances. True, and unrelated to the score. The score is a proxy for one narrow behaviour; it is not a measure of financial health. If you carry high-interest balances, paying them down is probably the best available use of money regardless of what any number does.

How to check, and how often

You are entitled to your credit reports, and the three bureaus must provide them. Checking through a bureau directly or an authorised service is free under U.S. law; some services that advertise free scores are providing a derived score from a single bureau rather than the reports themselves.

Reviewing annually is a reasonable cadence, and more often if something unexpected appears.

What to look for is the same on all three:

• Accounts you do not recognise
• Balances on cards you are not using
• Reported payments marked late that you paid on time
• Debts you have already settled that still appear as open
• Public records — judgments, liens, bankruptcies — that should or should not be present
• Accounts listed twice

Dispute errors in writing through the bureau, which will investigate. In the U.S. there is a right to free written dispute results, and a further right to the underlying documents after a denial.

The honest summary

A credit score is a compressed prediction about repayment behaviour, built by two competing formulas on two different data foundations, varying by bureau. It responds most strongly to payment history and reported utilization.

The levers worth pulling are modest and specific: report before you close, keep reported utilization low, dispute errors, do not open accounts you do not need, and do not close old ones. Beyond that, the score moves slowly, and it is worth remembering that it measures a small slice of what actually determines financial outcomes.