Amsterdam Menus · The Dossiers
The technical layer behind the capstone essay, one dossier per angle
The methods and evidence behind The tourist trap costs 85 cents: seven dossiers, each written as the investigation it was, with the stress tests visible. Corpus: Snapshot · June 2026, ~900 priced venues, roughly 4 in 10 of Amsterdam’s ~2,028 restaurants.
How the dossiers fit together. D1 gives every dish an address (the coordinate system). D3 prices each address: a typical and a 90% range. D4 splits all price variation into dish-type (59%), venue (17%) and item noise (24%); D3’s ranges are exactly D4’s non-dish 41%, displayed cell by cell (see D4, “the ladder and the rungs”). D5 shows the location slice of the venue part is real but small (+4.2%/SD), and D7 confirms it with a model-free commodity check. D6 splits the venue part further: two fanciness registers, with the premium concentrating where no reference price exists. D2 is the venue typology underneath, honest about being infrastructure. One corpus, one decomposition, seven views.
| # | Dossier | One-line punch |
|---|---|---|
| D1 | The dish coordinate system | Every dish gets an address; names lie, coordinates don’t |
| D2 | The venue typology | Infrastructure, not discovery, and honest about it |
| D3 | The going-rate atlas | Carpaccio is €15 in this town (€11.50–23, 90% of the time) |
| D4 | What moves the price | The plate is 59% of it; the room, 17%; the address, a rounding step |
| D5 | The tourist tax | Real, +4.2%/SD, and smaller than everyone believes |
| D6 | The fancy tax | The premium lives where there is no reference price |
| D7 | The Bitterballen Index | Order eight in the centre, pay for nine |