Editorial team

This site has one author, and that author also reviews what they write. This page says who that is, what the review checks, and what you can check for yourself.

Author and reviewer

Kostan Proff

Engineer by trade. Writes the guides, maintains the calculation engine, and checks every published figure against a primary source. The code that does the arithmetic is public, and the formulas are covered by unit tests that run on every change.

GitHub

What you can check

The arithmetic is public

The code behind every calculator is in a public repository, with the formulas it implements, the unit tests that check them against known published figures, and the checks that block a release with a broken page. You do not have to take the numbers on trust, and you do not have to ask.

Open the repository

Why there is only one name on this page

Plenty of sites list a writing team and a separate review team. This one does not, because there is not one. The person whose name is above writes the guides, checks the arithmetic, and publishes the result. Where a page says who wrote it, that is who wrote it.

A second name would be easy to invent and there is no honest way to do it. If a site says a reviewer checked something, the honest version of that claim is a person who did check it. So the site says what is actually true and backs it with something a reader can inspect: the source of every figure, the formula behind it, and the code that applies it, in public.

If an outside reviewer is added later, their name appears on the pages they actually reviewed and nowhere else.

What the review checks

Before a guide is published, the reviewer works through the same list every time:

  • The arithmetic. Every formula is checked against the source it comes from, and the worked example in the text is recalculated from scratch. Where a figure appears in the code, the two are compared.
  • The units. Percentages against decimals, months against years, dollars against thousands. Most mistakes in a financial document are a unit error.
  • The edge cases. A zero rate, a zero term, a negative value. Each is checked to produce a sensible answer rather than an error message.
  • The wording. The page must not tell the reader what to do. It explains what a number means and what it depends on, and stops there. A guide that recommends an action gets rewritten.
  • The sources. Each link is opened, checked for the claim it is attached to, and dated.

How drafting tools are used

Drafting tools help with the prose. They do not decide what a page says. Every figure, every formula and every claim is written or checked by a person, and no page goes live without the review above. A tool that produced a plausible number is exactly the kind of thing this process exists to catch.

The check that catches it is not only human. The calculation engine has unit tests that run on every change and block a release if a figure moves: the schedule has to sum to the amount borrowed, the loan has to close at exactly zero, and the period-by-period projection has to agree with the closed-form formula at every compounding frequency. A passage of prose cannot fail that way. A formula can, and when it does, the build stops.

Review dates

Every guide carries the date it was last reviewed. A guide whose content depends on outside data is reviewed at least twice a year; a guide that explains arithmetic is reviewed when the formula behind it changes, and in any case once a year. When a page is reviewed the date moves, whether or not the text changed, so the date tells you when a person last read the page.

Corrections

Corrections are made in the open. A page that has been corrected keeps its review date updated and the correction is described in thecontact thread it came from. A guide with a wrong figure is a defect in the review, not in the writing, and the review process is what changes.