Status: accepted

Accepted — merged with intent to implement.

RFC 0113: Define hospitality occupied room nights as paid room nights sold

Accepted by the UWMD maintainers, 2026-10-21, as drafted (owner: "RFC 0113 (#476) and RFC 0114 (#477) accepted as drafted."). Origin: issue #343 finding M07, left open by the metric variants table in spec/metrics/README.md ("whether to define occupied_room_nights as paid rooms sold"). It fixes the meaning of a stated field that adr and occupancy read, so it is normative. No formula changes.

Summary

Hospitality adr is rooms revenue ÷ rent_roll.occupied_room_nights, and occupancy is occupied_room_nights ÷ available_room_nights. The Format says only that occupied_room_nights is an integer ≥ 0, and its schema calls it "room nights sold". USALI and the OCC Comptroller's Handbook (Hotels) compute ADR and occupancy on paid rooms sold, excluding complimentary and house-use rooms. A producer that counts complimentary nights gets a lower ADR and a higher occupancy than the same hotel reported the standard way, and no check can tell. This RFC defines the field as paid room nights sold.

Motivation

• The same field, two numbers. Two producers stating the same hotel can disagree on ADR by the share of complimentary nights, and both records verify.
• Lenders read ADR on paid rooms. Underwriting packages and the OCC handbook use the USALI basis. A record that does not say which basis it used cannot be compared.

Proposed change

1. Format §4.3, hospitality variant. occupied_room_nights is the number of room nights sold for revenue in the period that rooms_revenue covers. Complimentary and house-use room nights are excluded. A room night in a package or contract whose price includes the room counts as sold (USALI, Rooms Department).
2. Optional complimentary_room_nights (D2). An integer ≥ 0. Producers MAY state the excluded nights separately, so the physical count stays recoverable. No metric reads it.
3. No formula change. The adr and occupancy dictionary entries, their round_to and their null guards are unchanged. The variants table records the basis as settled.
4. Schema. spec/schemas/rent-roll-hospitality.schema.json describes the field as paid room nights sold and admits complimentary_room_nights.

Compatibility analysis

No compatibility path (prelaunch; the prelaunch policy). The examples' stated counts are already paid counts; at implementation, each is checked against its narrative.

Conformance impact

No expected output changes. One new fixture states complimentary_room_nights beside a paid count, and shows ADR and occupancy computed on the paid count.

Reference implementation

The Format text, the schema fragment, the variants table row, the Python reader's schema copy if it has one, and the fixture.

Alternatives considered

• Leave the basis undefined and document it. This is today's state (#451). It leaves M07 open. Rejected.
• Define the field as physical occupied rooms, with a separate paid count. This inverts the industry basis for the field that ADR reads. Rejected.

Unresolved questions

• STR's benchmark treats some zero-rate contract nights differently from USALI's complimentary rooms (#343 M07). This draft follows USALI, which is what adr is compared against in the variants table. A separate STR-basis field is not proposed.

Decisions (accepted 2026-10-21)

#	Question	Proposal	Why / alternative
D1	Meaning of occupied_room_nights	Paid room nights sold; complimentary and house-use excluded	USALI and the OCC handbook. Alternative: physical occupied rooms
D2	Record the excluded nights	Optional complimentary_room_nights, read by no metric	Keeps the physical count recoverable. Alternative: no field
D3	Formulas	Unchanged	The fix is the input's meaning, not the arithmetic
D4	Versions	None allocated; prelaunch 4.x	RFC 0091 D6

Prior art

• Issue #343 (M07); spec/metrics/README.md "Variants in use" (#451).
• RFC 0103 (stated booked rooms revenue; CC-HOSP-05 cross-check).
• Draft RFC 0087 (hospitality revenue reconciliation basis), which this RFC neither depends on nor changes.