A credit score is produced by a proprietary statistical model that ingests dozens of variables from the bureau file and outputs a single number designed to rank-order borrowers by predicted risk. What it deliberately does not do is describe the file — it does not tell you which account is driving the score down, whether a negative mark is recent or years old and resolved, or whether a thin file is thin because the borrower is young or because they simply avoid credit.
This is by design, not a flaw — the score's entire value is in compressing complexity into something comparable across millions of files quickly. But that same compression is exactly why underwriting decisions are not, and should not be, made on the score number alone.
Consider two borrowers who both land at 680. The first has a short but spotless file: two credit cards, both paid on time every month, opened three years ago. The second has a longer file with one nine-month-old R9 collection that was paid and closed, sitting alongside several years of otherwise perfect payment history on multiple accounts. Both could plausibly produce a similar score, but they represent very different underwriting stories — one is a thin, clean file with limited data; the other is an established file with proven recovery from a specific past problem.
A lender who only sees '680' treats both borrowers identically. A lender, or broker, who reads the file itself can make a more informed judgment about which of the two represents the stronger long-term credit risk, and can prepare a submission note — the skill developed in Course 20 — that tells the underwriter which story applies before they have to dig for it themselves.
Bankruptcies, consumer proposals, judgments and collections — covered fully in Modules 07 and 08 — appear as discrete entries on the bureau file with dates, amounts and status, information a score number cannot convey on its own beyond its effect on the overall figure. The same is true of the inquiry list covered in Module 06: a lender reading the file can see exactly who pulled the borrower's credit and when, which is invisible from the score alone.
Address history, employer information as reported by creditors, and any consumer statement the borrower has added to explain a specific item also live only in the file itself. A borrower who added a written statement explaining a dispute or a specific hardship is giving the underwriter context that a plain score number could never carry.
Every module that follows this one is, in effect, teaching you to read one of these file-level details that the score cannot show: the utilization percentage behind a revolving trade line, the pattern in an inquiry list, the specific status and age of a collection or judgment, the province-specific retention rules governing how long a bankruptcy stays visible. None of this shows up if you stop at the score.
The practical habit this module is really arguing for is simple: pull the full file, not just the score summary, before forming a view on any borrower whose approval is not a foregone conclusion.
Why might two borrowers with the same credit score represent meaningfully different risk to a lender?
The score compresses many different underlying situations — file age, account mix, timing and severity of any past issues — into a single number, which means the same score can be reached from genuinely different files. The idea that identical scores mean identical files is exactly the misconception this module corrects. Scores are not random; they are a real statistical model, just not a substitute for reading the actual file. And plenty of factors beyond simple trade line count feed into both the score and the underlying story.
The intro and first module are free to read. Add your name and email once and the rest of this course opens — along with every other course on the site. No card, no trial.
Already unlocked on another device?
