ValueScreener

Lesson 1 of 8

What a model is for

A model establishes what must be true for a set of assumptions to give a result. It cannot tell you the assumptions are right, and it is not a forecast.

Javier Sanz, founder

By Javier Sanz

Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.

Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)

In one minute

  • A model turns a set of stated assumptions into a consistent set of statements, and its output is the consistency, not the number.
  • The three statements are joined at two points only: profit reaches the balance sheet through retained earnings, and cash reaches it from the bottom of the cash flow statement.
  • The useful question a model answers is what would have to be true, which is why the sensitivity lesson matters more than the base case.

A model does one thing well. Given a set of assumptions, it works out every consequence of them across three statements and refuses to let any of those consequences contradict another. That refusal is its whole value.

The two joins that make three statements one model

  1. 1Income statementRevenue down to net income
  2. 2Cash flow statementNet income, adjusted for non-cash items and working capital
  3. 3Balance sheetCash from the bottom of the cash flow statement; retained earnings from net income
  4. 4It balances, or an assumption is wrong
Illustrative round numbers, not any company's filings.

The two joins

Everything else in a model is arithmetic inside one statement. Only two lines cross between them.

Net income reaches the balance sheet through retained earnings: opening retained earnings plus net income less dividends equals closing retained earnings. And the closing cash balance at the bottom of the cash flow statement is the cash line on the balance sheet.

Get those two right and the balance sheet balances whenever the assumptions are internally consistent. When it does not balance, one of the assumptions has produced an asset or a liability that nothing paid for, and the failure is information.

What it establishes and what it cannot

A model establishes implications. If Costco opens twenty warehouses a year, holds its gross margin, and grows membership at the rate it has, then this is the operating income, this is the capital spending, this is the cash left over, and this is what the balance sheet looks like in five years.

Every one of those clauses is an assumption, and the model tests none of them. It cannot tell you whether twenty warehouses a year is possible, whether the margin holds, or whether members renew. What it can tell you is that those three assumptions together imply a specific amount of capital spending, and that if the capital spending is impossible then at least one of them is wrong.

Why it is not a forecast

A forecast is a claim about what will happen. A model is a claim about what follows from a set of premises. The two are confused constantly, and the confusion is what turns a model into a forecast wearing a decimal point.

The honest use runs the other way. Instead of asking what the model says the business is worth, ask what the model requires: what growth rate, what margin, what reinvestment, for the result to be a particular figure. That question has a checkable answer and the other one does not.

The rule this course follows

Every assumption is written where it can be changed, once, and referred to everywhere else. A number typed into two places is two assumptions that will disagree the first time one of them is edited, and a model that disagrees with itself is not an argument about anything.

Check yourself

Which two lines join the three statements?

  1. Revenue and operating income
  2. Net income, which reaches the balance sheet through retained earnings, and closing cash from the bottom of the cash flow statement
  3. Depreciation and capital expenditure

Show the answer

Answer: B.

Everything else is arithmetic within one statement. Those two lines are what make the three into one model, and what makes it balance or fail.

What does a model establish?

  1. What the business is worth
  2. The consequences of a stated set of assumptions, and whether those consequences are consistent with one another
  3. What management intends to do

Show the answer

Answer: B.

It works out every implication of the premises and refuses contradictions. It has nothing to say about whether the premises are true.

What is the honest way to use a model?

  1. Read off the value it produces
  2. Ask what growth, margin and reinvestment would have to be true for a given result, because that question is checkable
  3. Average several models together

Show the answer

Answer: B.

Running it backwards asks a question the evidence can answer. Running it forwards produces a number whose accuracy nobody can assess.

Sources

  • Costco Wholesale Corporation, Form 10-K for the fiscal year ended 31 August 2025, filed 2025-10-08, accession 0000909832-25-000101: the consolidated statements of income.
  • Item 7 of Form 10-K, management's discussion and analysis, where the operating drivers used in the following lessons are disclosed.

What this leaves out

  • The figure showing the two joins is illustrative and carries no company's numbers.
  • A model's consistency is not evidence that its assumptions are reasonable. A perfectly balanced model of an impossible business balances perfectly.
  • Costco appears throughout this course as a source of filed figures. Nothing here is a view on Costco or on any security, and no lesson in this course produces a price.

What the figure means, not what yours is

Lesson 1 of 8 of Building a model that balances. It teaches what this figure measures and where it stops meaning anything. It does not compute it for a company you hold, and a figure understood is not a figure checked.

Free audit

Research tool, not investment advice. See our methodology for how the product reads a filing and what it refuses to compute.

Once a week, what changed in the filings of US-listed companies, why it can matter to a thesis, and what deserves a closer review. Grounded in filings, not market noise.