Why a table beats a formula
Ask the internet how many BTUs a room needs and you will get a per-square-foot rule of thumb. Ask ENERGY STAR and you get a table — and the difference is not bureaucratic fussiness. The published room air conditioner sizing table runs fourteen rows, from 100–150 ft² at 5,000 BTU/h up to 2,000–2,500 ft² at 34,000 BTU/h, and the ratio between capacity and area changes as rooms grow. Worked out per square foot, the table spans roughly 20 to 34 BTU/h per ft²: small rooms need proportionally more capacity per square foot, and large rooms need less.
That curvature is why a flat multiplier misleads at both ends. Apply a mid-range per-square-foot figure to a 120 ft² home office and you will land under the smallest sensible unit’s duty; apply it to a 1,800 ft² open-plan space and you will shop for more capacity than the table calls for. The table already encodes the non-linearity, so the honest method is simple: measure the room, find its row, and only then think about adjustments.
Reading the table: ranges and boundaries
Measuring the room means floor area — length times width, in square feet. A 20 × 17 ft living room is 340 ft², which lands in the table’s 300–350 ft² row: 8,000 BTU/h. The only subtlety is what happens exactly at a published boundary. The rows behave as half-open ranges, so a boundary area starts the higher row: a room that measures exactly 150 ft² belongs to the 150–250 ft² row at 6,000 BTU/h, not the 100–150 ft² row below it.
It is worth being honest about what the table’s input does not include. The rows are keyed to floor area alone — the published method handles everything else through its three adjustments rather than through extra measurements. That makes the method fast, and it also defines its limits, which we will get to below.
The three published adjustments — and their order
ENERGY STAR publishes exactly three corrections to the table figure, and they come in a specific order. First, sun exposure: reduce the table capacity by 10% if the room is heavily shaded, or increase it by 10% if the room is very sunny. Second, occupancy: the table assumes two people, so add 600 BTU/h for each additional person who regularly occupies the room. Third, the kitchen: add 4,000 BTU/h if the unit will serve one.
The order matters because the adjustments are not all the same kind of arithmetic. The sun correction is a percentage of the base table capacity — it scales the room figure. The occupancy and kitchen corrections are fixed additions that stack on top of the scaled figure, because a body gives off roughly the same heat in a sunny room as a shaded one, and a stove is a stove. Take the sunny variant of our 340 ft² living room: 8,000 × 1.10 = 8,800 BTU/h. If that same room regularly held four people, you would then add 2 × 600 = 1,200 BTU/h on top — not 10% of anything.
Each correction traces to a real heat source: solar gain through the glazing, body heat from occupants beyond the assumed pair, and the appliance load of cooking. That is worth keeping in mind when you are tempted to skip them — they are not safety padding, they are loads.
Bigger is not better: what oversizing actually costs
The intuition that a larger air conditioner is a safer buy fails because a room air conditioner has two jobs, and only one of them is on the box. It removes heat, which is what the BTU rating measures — and it removes humidity, which it can only do while running. An oversized unit satisfies the thermostat quickly and shuts down before it has run long enough to wring the moisture out of the air. The result is the signature failure of oversized cooling: a room that is cold but damp and clammy. The frequent compressor start-stop cycles also waste energy and wear the machine.
ENERGY STAR is explicit that bigger is not better, and its guidance for the gaps between retail sizes follows directly: round down to the nearest thousand BTU rather than up. Window units come in steps — 5,000, 6,000, 8,000, 10,000, 12,000 and on up to 24,000 BTU/h at common retailers — so an adjusted target of 8,800 BTU/h sits between the 8,000 and 10,000 stock sizes. Modest undersizing costs a little pull-down speed on the hottest afternoons; oversizing costs dehumidification every single day. In a humid climate, the size below your target is usually the better pick; in a hot, dry one, the size above is a reasonable choice.
Window versus portable: mind the SACC
If you are comparing a window unit against a portable air conditioner, the headline BTU numbers are not speaking the same language. A portable unit sits entirely inside the room and pushes its hot exhaust out through a duct — and that duct leaks heat back into the very space being cooled. The U.S. Department of Energy’s SACC rating (seasonally adjusted cooling capacity) exists precisely to account for this, which is why a portable’s traditional ASHRAE BTU figure overstates its delivered cooling compared with a window unit carrying the same number.
The room-load side of the method is unaffected — your room needs what it needs regardless of which machine supplies it. So size the load from the table and adjustments as usual, then compare portable units by their SACC rating, not the larger ASHRAE number on the same box. Comparing a window unit’s rating against a portable’s ASHRAE figure will systematically flatter the portable.
When a room-unit table is the wrong tool
The published table ends at 2,500 ft², and well before that ceiling a single window unit starts fighting physics: it cools the room it sits in and leaks capacity through every open doorway. Multi-room and whole-house cooling are not bigger versions of the same problem — they are a different problem, governed by insulation levels, window area and orientation, air infiltration, and duct or unit placement, none of which the room table sees.
That is the boundary where a lookup method should hand off to a real load calculation. A Manual-J-style calculation builds the sensible and latent loads from the building itself and returns equipment tonnage, which is what you actually need for whole-house decisions — our HVAC load calculator does exactly this. For a single room with a window unit, though, the ENERGY STAR method remains the published, verifiable path: floor area, table row, three adjustments in order, then round down to the nearest stock size. Our window AC BTU calculator walks those steps with the fourteen-row table built in.