How Credit Scores Are Actually Calculated

You applied for something, a card, an auto loan, an apartment, and the number that came back wasn't what you expected. Now you're trying to figure out what actually moves that number before your next application. Here's the breakdown, in order of weight.

Your FICO score is built from five pieces: payment history (35%), amounts owed (30%), length of credit history (15%), new credit (10%), and credit mix (10%). Payment history and what you owe make up almost two thirds of the score. Everything else is fine tuning. If you remember nothing else, remember that.

What actually moves the number

Payment history, 35%. This is whether you've paid on time. Not how much you paid, just whether the payment showed up by the due date. One payment 30 days late can knock 60 to 100 points off a score in the 700s. A payment 90 days late is worse, and it stays on your report for seven years, though its damage fades as it ages.

Amounts owed, 30%. This is mostly your credit utilization, the balance on your revolving accounts (mainly credit cards) divided by your total credit limit. Say you have three cards with a combined limit of $15,000, and you're carrying $6,000 across them. That's 40% utilization. Scoring models start penalizing you noticeably above 30%, and the penalty gets sharper above 50%.

Here's the part people miss: utilization is calculated from whatever balance gets reported to the bureaus, which is usually your statement balance, not your balance today. You can pay a card off in full every month and still show 40% utilization if you spent heavily right before the statement closed.

Length of credit history, 15%. This includes the age of your oldest account, your newest account, and the average age across all of them. There's no shortcut here except time. This is the real argument against closing your oldest card even if you never use it.

New credit, 10%. Every hard inquiry, an application for a card or loan, takes a few points off, usually 5 to 10, and stays on your report for two years. One inquiry is minor. Five inquiries in three months for unrelated credit tells the model you might be in financial trouble, and it reacts accordingly.

Credit mix, 10%. Having both revolving credit (cards) and installment credit (auto loan, mortgage, student loan) helps a little. This is the smallest factor and not worth manufacturing debt over.

A worked example

Take someone with a 680 score. They have two cards: a $3,000 limit with a $2,700 balance, and a $7,000 limit with a $500 balance. Total utilization: $3,200 owed against $10,000 available, 32%. They pay both on time every month, no late payments ever, credit history averaging six years.

Their biggest lever isn't payment history, that's already maxed out. It's the $2,700 balance sitting at 90% utilization on that smaller card. Scoring models look at utilization both in aggregate and per card, so a single maxed out card drags the score down even if the overall number looks moderate.

If they pay that card down to $600, utilization on it drops to 20%, and aggregate utilization drops to $1,100 over $10,000, 11%. That single move, no new accounts, no closed accounts, nothing else changed, can realistically move a score in the 680 range up 20 to 40 points within one billing cycle, because utilization is reported monthly and reacts fast.

Compare that to opening a new card to "improve credit mix." That triggers a hard inquiry (small hit), lowers average account age (small hit), and adds a $0 balance card (which can help utilization later, but takes a cycle to show up, and only if you don't spend on it). For someone trying to fix a score fast, paying down an existing balance almost always outperforms opening something new.

What people get wrong

The biggest mistake is closing a card to "simplify," usually the oldest one or one with a low limit. Closing it doesn't erase the history immediately, but it removes that limit from your utilization math going forward, and eventually the age stops counting toward your average once it drops off. If you have a card you don't use, put one small recurring charge on it, a streaming subscription, and let it sit paid off. That keeps it open and reporting without any downside.

The second mistake is thinking a credit score is one number. You have dozens. FICO alone has multiple versions (FICO 8, FICO 9, industry-specific versions for auto and mortgage lending), and VantageScore is a separate model entirely with its own weighting. The number you see on a free app is usually VantageScore 3.0, and it can run 20 to 40 points different from the FICO score a mortgage lender actually pulls. Don't panic over a 15 point swing between two apps. They're not measuring the exact same thing.

The third mistake is checking your own score and thinking it dinged you. Checking your own credit is a soft inquiry. It does not affect your score, no matter how many times you do it. Only inquiries from lenders you've applied with count.

The honest limitation

I can walk you through the math, but I can't tell you your exact score movement in advance, and neither can anyone else. FICO doesn't publish the exact formula, only the weighted categories. Two people with identical utilization and payment history can score differently based on account age, inquiry timing, and details in the model that aren't public. Anyone who tells you "do X and you'll gain exactly Y points" is guessing with more confidence than the data supports. What's reliable is direction: pay down revolving balances, never miss a due date, and let accounts age. The exact number is noise. The direction is not.

If you're trying to fix a score because you're staring down a real decision, a mortgage application, a car loan, a lease, this is one piece of a bigger picture that includes your income, your debt load, and what you're actually trying to qualify for. That's what Foundation walks through. You can find it at readmoneydecoded.com/foundation.

← All articles