How Credit Scores Work: The Complete Scoring Algorithm Breakdown (2026)

Here's a number that controls more of your financial life than your salary: your credit score. It determines whether you get approved for a mortgage, what interest rate you pay on a car loan, and — increasingly in 2026 — whether you land that apartment lease or even that job offer.

Yet most people have no idea how the number is actually calculated. They just know they're supposed to "keep it high."

We're going to fix that. At ScoreNerds, we don't do vague. We break down exactly how FICO and VantageScore algorithms process your credit data, what weight each factor carries, and what the scoring models actually reward. The data shows that understanding these mechanics is the fastest path to improving your score — because you stop guessing and start optimizing.

Let's decode the algorithm.

The Basics: What a Credit Score Actually Is

A credit score is a three-digit number — typically between 300 and 850 — generated by a mathematical algorithm that predicts the probability you'll become 90+ days delinquent on any credit obligation within the next 24 months. That's it. It's a risk prediction tool.

The algorithm ingests your raw credit report data (tradelines, inquiries, public records), runs it through a scoring model, and outputs a number. Higher number = lower predicted risk of default.

There are two dominant scoring companies:

• FICO (Fair Isaac Corporation) — Used in approximately 90% of U.S. lending decisions, according to FICO's own disclosure. Founded in 1956, they've been refining their models for nearly 70 years.
• VantageScore — Created in 2006 by the three major credit bureaus (Experian, Equifax, TransUnion) as a competitor. Used heavily in credit monitoring apps and increasingly by lenders. VantageScore reports that over 2,500 lenders used VantageScore 4.0 as of 2024.

Important distinction: your credit score is calculated fresh every time it's requested. There's no stored "score" sitting in a database. When a lender pulls your credit, the bureau runs your current data through the scoring algorithm and generates a score at that moment. This is why your score can fluctuate day to day as new data is reported.

The Five FICO Scoring Factors (With Exact Weights)

FICO has publicly disclosed the five categories of data their algorithm evaluates, along with approximate percentage weights. These weights can shift slightly depending on your individual credit profile — for example, someone with a thin file may see different factor emphasis than someone with 20 years of history — but the baseline weights are consistent across FICO Score 8, the most widely used version in 2026.

Let's break down exactly how each factor works inside the algorithm.

For a deep dive into each factor with optimization strategies, see our complete guide to the five credit score factors.

Factor 1: Payment History (35%)

This is the single heaviest-weighted factor — and the data shows it's also the most punishing. According to FICO's research, a single 30-day late payment can drop a 780 score by 90-110 points. For someone sitting at 680, the same late payment might only cost 60-80 points. The algorithm penalizes higher scores more severely because the deviation from expected behavior is larger.

The algorithm evaluates payment history across several dimensions:

• Severity: 30 days late is bad. 60 days is worse. 90+ days, charge-offs, and collections are devastating.
• Recency: A late payment from 6 months ago hurts far more than one from 5 years ago. The impact decays over time but remains on your report for 7 years.
• Frequency: One isolated late payment is treated very differently than a pattern of delinquency.
• Amount: Being late on a $10,000 balance is weighted more heavily than being late on a $200 store card.

ScoreNerds Data Point: According to Experian's 2025 consumer credit review, only 1.4% of consumers with scores above 800 have any late payment on their record within the past 7 years. Meanwhile, a single 30-day late payment can drop a 780 score by 90-110 points — but the same late payment on a 680 score costs only 60-80 points. The algorithm penalizes deviations from established patterns more severely at higher scores.

Payment history isn't just important — it's effectively a prerequisite for top-tier scores.

Factor 2: Amounts Owed / Credit Utilization (30%)

This is where the algorithm gets interesting. "Amounts owed" is really about utilization ratios — how much of your available revolving credit you're currently using.

The formula is straightforward: (Total revolving balances) / (Total revolving credit limits) = Utilization ratio

But here's what most people miss: the algorithm evaluates utilization at both the aggregate level AND the individual card level. You could have a 10% overall utilization but if one card is maxed at 95%, the algorithm flags that individual card's ratio as a negative signal.

ScoreNerds Data Point: FICO's own published research shows consumers with scores above 785 carry an average utilization of just 7%. Not under 30% — 7%. The "keep it under 30%" advice that dominates personal finance content is the minimum, not the optimum. Our credit score experiments confirm the biggest score jumps happen when utilization drops below 10%.

The algorithm also looks at:

• Installment loan balances vs. original amounts: Paying down a $25,000 car loan shows positive repayment behavior.
• Number of accounts with balances: Carrying balances on 8 cards reads differently than carrying the same total across 2 cards.
• Total balance trajectory: Are your total balances trending up or down over recent months?

Factor 3: Length of Credit History (15%)

The algorithm rewards patience. It evaluates three metrics:

1. Age of oldest account — How long you've been in the credit system
2. Average age of all accounts — The mean age across your entire profile
3. Age of newest account — How recently you opened a new account

According to Experian data, consumers with FICO scores above 800 have an average credit history length of 25+ years. This is the one factor you genuinely cannot hack — it just takes time. But understanding how it works prevents you from accidentally destroying it (like closing your oldest credit card).

Factor 4: Credit Mix (10%)

FICO's algorithm gives a modest boost for demonstrating that you can manage different types of credit responsibly. The main categories:

• Revolving credit: Credit cards, HELOCs, retail store cards
• Installment loans: Auto loans, student loans, personal loans
• Mortgage debt: Treated as its own category in the algorithm

At just 10% weight, this is not a factor worth manufacturing accounts for. Don't take out a car loan to "improve your credit mix." But if you're comparing two consumers with otherwise identical profiles, the one with a mortgage + credit cards + an installment loan will score higher than the one with only credit cards.

Factor 5: New Credit (10%)

Every time you apply for credit and the lender performs a hard inquiry, the algorithm registers it. FICO's data shows the average hard inquiry impact is fewer than 5 points — but multiple inquiries in a short window (outside of rate-shopping) can signal financial distress.

The algorithm includes an important safeguard: rate-shopping protection. If you're shopping for a mortgage or auto loan and multiple lenders pull your credit within a 45-day window (FICO Score 8) or 14-day window (older FICO versions), all those inquiries count as a single inquiry. The algorithm assumes — correctly — that you're comparison shopping, not seeking multiple new loans.

Hard inquiries remain on your report for 2 years but only affect your score for 12 months. For a complete breakdown of which actions trigger hard pulls versus soft pulls — including common surprises like rental applications and utility signups — see our hard inquiry vs. soft inquiry guide.

How the Algorithm Actually Processes Your Data

The scoring algorithm doesn't just add up percentages. It uses a scorecard-based system — think of it as a decision tree. When your credit data enters the FICO model, the algorithm first assigns you to a "scorecard" (a subpopulation of consumers with similar profiles).

For example, consumers with a bankruptcy on their record are scored on a different scorecard than consumers with no derogatory marks. This is why two people with the same utilization ratio and payment history can get different scores — they may be on different scorecards where the relative value of each factor differs.

The process works like this:

1. Data ingestion: The bureau provides your raw tradeline data to the scoring model
2. Scorecard assignment: The algorithm places you into the appropriate scorecard subgroup
3. Factor evaluation: Each of the five factors is scored against others in your scorecard group
4. Score output: The weighted sum produces a number between 300 and 850

This scorecard system is why the factor weights are "approximate" — they shift depending on which scorecard you land on. A thin-file consumer on the "limited history" scorecard may see credit mix weighted more heavily than 10%, while a seasoned borrower's scorecard may weight payment history slightly differently.

FICO Score Ranges: What the Numbers Mean

Not all 550-point ranges are created equal. Here's how lenders, financial institutions, and the industry at large categorize FICO score ranges in 2026:

The Federal Reserve's 2025 Survey of Consumer Finances found that the median U.S. credit score reached 715, continuing a steady upward trend. But that headline number obscures massive demographic gaps — see our average credit score by age breakdown for the full picture.

For a detailed analysis of what counts as a good score in today's lending environment, check our what's a good credit score in 2026 guide.

What Lenders Actually See When They Pull Your Credit

When a lender pulls your credit for a lending decision, they don't just get a single number. They receive:

• The score itself — from whichever model version they've contracted with the bureau to use
• Reason codes (up to 4) — specific explanations for why your score isn't higher (e.g., "proportion of balances to credit limits is too high," "length of time accounts have been established")
• The full credit report — every tradeline, inquiry, public record, and collection
• Score version identifier — which FICO or VantageScore model generated the number

Here's a data point most consumers don't know: different lenders use different FICO versions for different products. The most common versions in use as of 2026:

This is why the score you see on Credit Karma (VantageScore 3.0) can differ significantly from the score a mortgage lender pulls (FICO Score 10T). They're literally different algorithms running against potentially different bureau data.

FICO vs. VantageScore: Key Algorithm Differences

Both models score from 300-850 and use similar data, but the algorithms differ in important ways. Understanding these differences matters because the score your banking app shows you (usually VantageScore) may not match what a lender uses (usually FICO).

The practical impact: VantageScore tends to be more forgiving for consumers with limited or recovering credit, while FICO's models tend to produce higher scores for consumers with long, clean payment histories. A 30-50 point difference between your FICO and VantageScore is completely normal — don't panic if Credit Karma shows something different from your lender's number.

We built an entire comparison guide on this: FICO vs. VantageScore — which score matters more.

Common Credit Score Misconceptions (Debunked With Data)

The internet is full of credit score advice that ranges from "technically true but misleading" to "completely wrong." Let's correct the biggest ones using actual data. For the full rundown, our 12 credit score myths debunked with real data digs deeper into each of these and more.

Misconception 1: "Carrying a small balance helps your score"

The data says: No. This is perhaps the most persistent myth in consumer finance. FICO has explicitly stated that carrying a balance does not help your score. The algorithm compares your statement balance to your credit limit — and a $0 statement balance produces the best possible utilization signal. Experian's data confirms that consumers with 0% utilization do score slightly lower than those with 1-3%, but that's because $0 balances often indicate inactive accounts, not because carrying a balance helps. The solution: use your cards and pay in full before the statement closes.

Misconception 2: "Closing old credit cards is always bad"

The data says: It depends. Closing a card affects two factors: credit utilization (your total available credit shrinks, pushing utilization up) and average age of accounts. However, the closed account remains on your report and continues aging for 10 years after closure. So the average-age impact is delayed. If the card has an annual fee you're not getting value from, closing it may be the right call — especially if you have plenty of remaining available credit to absorb the utilization change.

Misconception 3: "Your income affects your credit score"

The data says: Absolutely not. Your credit report contains zero income information. The algorithm doesn't know if you earn $30,000 or $3,000,000. Income affects your debt-to-income ratio, which lenders evaluate separately from your credit score during underwriting. But the score itself? It's purely based on how you manage credit, not how much you earn.

Misconception 4: "Checking your own score hurts it"

The data says: No. Self-checks are soft inquiries and have exactly zero scoring impact. The CFPB has repeatedly confirmed this in consumer education materials. Only hard inquiries from lender-initiated credit checks affect your score, and even those average fewer than 5 points of impact according to FICO research.

Misconception 5: "You need to be in debt to have a good score"

The data says: Misleading. You need active credit accounts — not debt. A consumer who uses one credit card for everyday purchases and pays it in full monthly will build an excellent score. You don't need a mortgage, a car loan, or credit card debt. You need demonstrated, responsible use of credit facilities over time.

Lender Approval Thresholds in 2026

While every lender has proprietary underwriting criteria beyond just the score, the data shows clear threshold patterns across the industry. Based on aggregated 2025-2026 lending data:

Generational Score Gap (2026): Gen Z's average credit score dropped to 676 — the lowest of any generation — while Americans aged 78+ maintain an average of 760, according to LendEDU's 2026 generational credit analysis. The 84-point gap is largely explained by credit history length (15% of FICO) and utilization patterns (30% of FICO). Understanding the algorithm is how younger consumers close this gap faster.

The math is clear: the cost of a mediocre credit score compounds over a lifetime. A Federal Reserve study estimated that consumers with subprime scores pay an average of $200,000 more in interest over their lifetimes compared to consumers with excellent credit.

See which credit cards you can realistically get approved for based on your score range: best credit cards by score bracket.

What Changed in 2026: FICO 10T and the Trended Data Era

The biggest shift in credit scoring for 2026 is the evolving rollout of FICO Score 10T and FICO Score 10. In July 2025, FHFA Director Pulte announced that lenders can choose between Classic FICO and VantageScore 4.0 for Fannie Mae and Freddie Mac loans — reversing the earlier bi-merge mandate and maintaining tri-merge reports. Full FICO 10T adoption is now expected by Q4 2026, with credit bureaus delivering historical FICO 10T data to the GSEs for validation.

The "T" stands for "trended data." Previous FICO models took a snapshot of your credit profile at a single point in time. FICO 10T analyzes 24 months of payment behavior patterns. This means the algorithm can now distinguish between:

• Transactors (pay in full monthly) — rewarded
• Revolvers (carry balances month to month) — penalized relative to transactors
• Improving trajectories (balances trending down) — given credit for positive momentum
• Deteriorating trajectories (balances trending up) — flagged as increasing risk

According to FICO's published research, FICO 10T produces scores that shift by more than 20 points for approximately 40 million Americans compared to FICO Score 8. Consumers who consistently pay in full will see score increases. Consumers who have been carrying growing balances will see declines.

The takeaway for 2026: the algorithm now rewards behavioral patterns, not just current-state snapshots. Paying down debt consistently — even slowly — will show up as a positive signal that previous model versions missed.

New in 2026: BNPL, Medical Debt, and Alternative Data

Three additional developments are reshaping what the scoring algorithm sees in 2026:

Buy Now, Pay Later (BNPL) Enters the Algorithm

FICO launched dedicated BNPL scoring models — FICO Score 10 BNPL and FICO Score 10T BNPL — in late 2025. For the first time, BNPL payment history from platforms like Affirm and Klarna feeds into credit score calculations. The algorithm treats BNPL differently from traditional credit: multiple BNPL loans are grouped together rather than each counting as a separate new account. FICO's early testing found that consumers with five or more Affirm loans typically saw scores increase or hold steady — as long as payments were on time. For a full breakdown of which BNPL providers report to which bureaus and the real scoring consequences, see our guide to how Buy Now Pay Later affects your credit score.

ScoreNerds Data Point: FICO's simulations show most BNPL users experience a score change of approximately ±10 points when BNPL data is incorporated — similar to opening a single new traditional account. For the estimated 45 million Americans using BNPL services regularly, this means payment behavior that was previously invisible to the algorithm now counts.

Medical Debt Removal

The CFPB's rule eliminating most medical debt from credit reports took full effect in 2025, following the bureaus' earlier removal of paid medical collections and debts under $500. According to CFPB estimates, approximately 15 million Americans saw medical collections removed from their reports — producing score increases of 20-40 points for affected consumers. The algorithm now treats medical debt as a fundamentally different risk signal than other types of collections.

Alternative Data: Rent, Utilities, and Streaming Payments

VantageScore 4.0 natively incorporates rent and utility payment data. FICO now supports it through Experian Boost and third-party rent reporters like Rental Kharma and RentReporters. Experian reports that Boost users see an average FICO score increase of 13 points. For thin-file consumers — particularly younger adults building credit — alternative data can be the difference between having a scoreable file and being "credit invisible."

How to Actually Improve Your Score (The Data-Driven Way)

Now that you understand the algorithm, the optimization strategy becomes obvious. In order of impact:

1. Never miss a payment. Set up autopay for at least the minimum on every account. This protects the 35% factor. One late payment can undo years of positive history.
2. Crush your utilization. Get below 10% — ideally below 5%. This is the fastest lever because utilization has no memory. Drop your balances and your score responds within one reporting cycle. The 30% factor responds immediately.
3. Stop opening unnecessary accounts. Every new account drops your average age and adds a hard inquiry. Only open new credit when there's a genuine financial benefit.
4. Keep old accounts open. Even if you don't use a card, keep it open if there's no annual fee. It contributes to your available credit (lowers utilization) and your average account age.
5. Diversify over time (don't force it). If you naturally need an installment loan, it helps your mix. But never take on debt for the sake of credit mix.

We maintain a comprehensive, data-backed optimization guide: how to improve your credit score — including specific strategies we've tested in our score experiments.

Frequently Asked Questions

How often does my credit score update?

Your credit score recalculates every time a lender or service pulls it — it's not a stored number. Most creditors report to the bureaus every 30-45 days, so meaningful changes typically appear within that window. However, some issuers report more frequently, and certain events like paying off a balance can reflect within days if your issuer does mid-cycle reporting. We break down every detail — including creditor reporting schedules, rapid rescoring for mortgage applicants, and FICO vs. VantageScore update differences — in our dedicated score update frequency guide.

Do I have one credit score or multiple?

You have dozens. FICO alone has over 60 industry-specific scoring models (FICO Auto Score, FICO Bankcard Score, etc.), and each of the three bureaus may have slightly different data. Add VantageScore versions and you could have 100+ scores. The version your lender uses depends on the credit product you're applying for.

Does checking my own credit score lower it?

No. Checking your own score is a "soft inquiry" and has zero impact on your score. Hard inquiries — which happen when a lender checks your credit for a lending decision — can lower your score by 5-10 points temporarily. FICO data shows the average hard inquiry impact is less than 5 points for most consumers.

What is the minimum data needed to generate a credit score?

For FICO, you need at least one account that's been open for 6 months or longer, and at least one account reported to the bureau within the last 6 months. VantageScore is less restrictive — it can score consumers with just one month of history and one account reported within the past 24 months. This difference means VantageScore can score approximately 37 million more Americans than FICO.

Why is my FICO score different from my VantageScore?

FICO and VantageScore use different algorithms with different factor weights. For example, VantageScore weighs your total credit usage and balances as "extremely influential" while FICO assigns payment history the highest weight at 35%. They also treat collections, authorized user accounts, and thin files differently. A 30-50 point difference between the two models is completely normal.

The Bottom Line

Your credit score isn't mysterious — it's math. The FICO algorithm takes your tradeline data, sorts you into a scorecard, evaluates five weighted factors, and outputs a number between 300 and 850. The weights are public. The mechanics are documented. The data on what actually moves scores is available to anyone willing to dig.

The problem has never been that credit scoring is "too complex." The problem is that most credit score advice stops at "pay your bills on time" and never explains the actual algorithm. Now you know better.

At ScoreNerds, we'll keep decoding the data so you don't have to guess. Start with understanding where your score falls in the ranges, learn the five factors in detail, then build a strategy to improve your score based on what the algorithm actually rewards.

The data doesn't lie. Your credit score is just waiting for you to optimize it.

Back to the Credit Scores hub for more data-driven guides.