Sources

Where every number came from.

And, more usefully, which ones we made up.

A game that teaches with numbers has one obligation above all the others: to be clear about which numbers are recorded history and which are ours. Mixing them is how a teaching game quietly becomes a confident-sounding lie.

So every figure in Kitchen Table belongs to exactly one of three classes, and the class travels with the number. Where you see a small marker beside a figure in the game, it is one of these:

[A] RECORDED A published series or a law. Somebody else collected it, we link to them, and you can check it without taking our word for anything. The Fed rate path, inflation, the mortgage average, Regulation Q.

[R] REPRESENTATIVE Typical of the period, or derived from something that is — but not a single recorded quote. There was never one savings rate or one credit-card rate; there were thousands. These figures are honest about the shape of the era without pretending to be a receipt from a named bank on a named day.

[B] AUTHORED We invented it, to be plausible for the era. The families, their wages, their house, their car, the grocery bill. This is the part of the game we wrote.

The rule we hold ourselves to: a number may never be promoted to a class it has not earned. This page exists because we got it wrong once — the footer used to claim the prices "are not invented" while showing an authored $58,000 house. An [A] is only worth something if we are strict about the [R]s and [B]s.

The register

Rendered from js/sources.js, which is the same register the game itself cites from — so this page cannot drift out of date with the game.

“Checked” means we re-ran the number. A class is a claim about where a figure came from, and a claim nobody ever re-tests rots quietly. tools/verify-rates.js pulls each cited series from FRED over that scenario's own window and prints our figure beside the published one; what it found is recorded below, including where we are off. An entry marked [A] with no check is an assertion, not a citation.

The ideas, and where they come from

The game also makes claims — about how households behave, why a cash buffer is worth holding, what a fixed payment does in an inflation. Those are a different kind of thing from a number, and they get their own four classes, because "a result in the economics literature" and "a framework from a bestselling personal-finance author" are not the same promise and should never look the same on a page.

[A] ACADEMIC A result from primary research or official data. Friedman on permanent income, Kahneman and Tversky on reference points, Carroll on buffer saving.

[S] SYNTHESIS From a credible synthesis or academic book rather than a single paper — Shiller on narratives, Mullainathan and Shafir on scarcity.

[P] PRACTITIONER A useful framework from a popular-finance writer. An argument with a name on it, not a finding. We use these and we label them, and we never encode one as a law.

[N] OURS Our own interpretation, written for this game.

What we will not do with these: turn a practitioner's rule into an objective truth, quote a lab coefficient as a constant of nature, or let a principle appear before you have done something that demonstrates it. The game says one line during a quarter and saves the explanation for the end — see js/principles.js, where each idea also carries the popular misreading of it that we are trying not to teach.

On the model

The economy is not scripted and the outcomes are not authored. Both scenarios are driven by the real quarterly Fed funds path and run through the same engine as THE CHAIR: an expectations-augmented Phillips curve, Okun's law, policy transmission lags and a credibility channel.

It is calibrated against published data rather than tuned by eye, and here is the scoring, so you can see how far off it is:

Autumn 1979 — inflationmean abs. error 1.35pp

Autumn 1979 — unemployment0.56pp

The Hot Summer — inflation0.38pp

The Hot Summer — unemployment0.25pp

Errors against the FRED/BLS actuals, quarter by quarter, over each eight-quarter window. A model this size will never be exact and one that claimed to be would be lying. Publishing the error is the point.

It is still a model. It is a simplification of an economy, which is a thing nobody has ever successfully modelled. Nothing in this game is a forecast, and nothing in it is advice — see the terms.

Back to the game · Privacy · Terms