Yield Simulation

A pro forma is a low-resolution simulation of occupancy: rent, vacancy, and expenses summarize assumptions about people using a building. A higher-resolution simulation could represent individual choices, such as whether a resident uses a space comfortably or avoids it because of a design condition.

What fits, who stays

One floor of a 160 ft double-loaded bar, drawn at one scale in every example: 30 ft deep units either side of a 6 ft corridor, and a 24 ft core with the elevator on one side. Studios are 14 ft wide (420 sf), 1 BRs 22 ft (660 sf) and 2 BRs 32 ft (960 sf), asking $1,400, $1,900 and $2,600 a month. Units fill each run of wall in order; one that does not fit starts the next run, and the rest of the run is left over. The flat pro forma lets every unit 95% of the year. Unit by unit, each starts at 95% and loses 6 points if it shares a wall with the core and 2 points for every full 40 ft between its door and the elevator door. Bars show the rent each unit loses to vacancy; a full bar is $6,000 a year. Sizes, rents and penalties are illustrative assumptions, not market or behavioral data. Egress, structure and code are not modeled.

Test fits simulate a different thing: geometric yield. Given an envelope and program, they explore how many spaces fit and how area is used. The two models answer related but distinct questions—what can fit, and what people may choose to occupy. Linking design assumptions to later operating evidence could improve future forecasts, but only if actual results are measured and compared.

One floor, two simulations

The study draws one floor of a 160 ft apartment bar: 30 ft deep units either side of a 6 ft corridor, and a 24 ft core with the elevator on one side. Studios are 14 ft wide, one-bedrooms 22 ft and two-bedrooms 32 ft. The test fit lines units up along each run of wall until the next one no longer fits. With one-bedrooms only and the core at one end, 13 units fit, 8,580 sf of rentable area, with 10 ft of wall left over. A flat 5% vacancy collects $281,580 a year, and the bars beside the plan show every unit losing the same $1,140.

Unit by unit, every door gets its own resident instead of a share of an average. The study gives them two assumed dislikes: a shared wall with the core costs 6 points of occupancy, and every full 40 ft between the elevator door and the unit door costs 2. The same plan now collects $272,004 a year, $9,576 less. 11 of the 13 residents are less likely to stay; the three worst-placed units lose $2,508 a year each, the two nearest the elevator $1,140.

Move the core to the middle. The test fit does not notice: the same 13 one-bedrooms fit, and the flat pro forma still collects $281,580. Unit by unit, the longest walk falls from more than 120 ft to under 80 ft, but two units now share the core’s walls. The floor collects $275,196, $3,192 more than with the core at the end. Only the finer model can tell the two plans apart.

The test fit has its own resolution. Slide the core to 12 ft from the end and the run beside it is too short for a 22 ft unit: 12 one-bedrooms fit, with 32 ft of wall left over. The flat pro forma falls to $259,920 and the unit-by-unit model to $253,080. Losing one unit costs $21,660 a year at the flat rate, more than six times what moving the core to the middle gained.

Make every unit a studio, with the core in the middle, and 19 studios fit, with 30 ft of wall left over. The flat pro forma rises to $303,240 and the unit-by-unit model to $297,864. Whether 19 studio households want this floor is in neither model: rents are fixed inputs here, and demand for a unit type is a market question.

Put the core back at the end and a 20th studio fits, with 16 ft of wall left over. The walks are longer and 15 of the 20 residents are less likely to stay, yet the floor collects more: $319,200 flat and $309,792 unit by unit, against $297,864 with the core in the middle. With studios, both models prefer the extra unit.

The sizes, rents and dislikes are the study’s assumptions, not measured behavior. Leasing and turnover recorded unit by unit are what would test them.

Neither simulation predicts behavior or returns with certainty. Their value is making assumptions visible at the resolution the decision requires.