The 7-day test: nights ÷ reservations, not nights ÷ 365
By Max Medvedev · 24 Jul 2026 · updated 6 Aug 2026
The formula, exactly
Rental losses are passive by default, and passive losses cannot touch W-2 income. The way out for a short-term rental — the front gate of what forums call "the STR loophole" — starts with one number: if the average period of customer use is seven days or less, the activity is not treated as a rental activity at all under Reg. §1.469-1T(e)(3)(ii)(A).
That average is a division with exactly two inputs:
Average stay = total guest nights ÷ number of reservations.
Rent 180 nights across 50 reservations and your average stay is 180 ÷ 50 = 3.6 nights — inside the seven-day line. Nothing about the calendar year, days listed, or occupancy enters the math.
The mistake that flips verdicts
The error we see most is dividing by the days in the year, or by the days the property was available — occupancy math wearing the test's clothes:
| The math | Result | What it means |
|---|---|---|
| 180 nights ÷ 50 reservations | 3.6 | The regulation's formula — inside the line |
| 110 nights ÷ 13 reservations | 8.5 | Over the line — the exception is gone |
| 110 nights ÷ 365 days | 0.3 | Occupancy math — says nothing about the test |
The second and third rows are the same property: ten weekend stays plus three month-long winter bookings. Divide by 365 and it "passes" comfortably; divide by reservations — the actual test — and it fails. A few long stays drag the average up faster than most owners expect, so check the mix before accepting that 30-night booking.
What passing actually gets you
Less than people think, and this is the part worth being precise about. An average stay of seven days or less only removes the automatic passive label. It does not create a deduction, and it does not make your loss non-passive by itself — you still have to materially participate, usually by winning the 100-hour race against your cleaner.
This is one of four gates, and clearing it early tells you nothing about the other three. The full qualification path walks them in order.
Mirch v. Commissioner, T.C. Memo 2025-128, is the clean warning: the court agreed the stays averaged seven days or less, and the owners still lost — the hours were not there. The average-stay math opens the door; the participation log walks through it.
The 30-day route, briefly
There is a second door: an average stay of 30 days or less combined with significant personal services — staff-provided, hotel-adjacent attention during the stay, well beyond cleaning between guests. Most self-managed properties do not clear that bar, which is why the seven-day route is the one that matters in practice.
Personal use can void all of it
One habit undoes both routes: using the place yourself. If personal use exceeds the greater of 14 days or 10% of the days rented at fair value, the dwelling becomes a residence under IRC §280A(d)(1) — and §280A(c)(5) then caps deductions at rental income. No hours log fixes that; the loss simply cannot exist that year, and the disallowed amount carries forward against future rental income.
At 180 rented days the limit is 18 personal days. Family stays count even at full price, and friends renting below market count too. Track those nights with the same care as the hour log.
The whole path, computed
Here is a year where the math holds together — the average stay, the hour race, the depreciation, and what is left of the W-2, computed live by the engine:
Every figure above traces to a rule and a citation. That is what "defensible" means here: a record where each number can answer for itself when someone asks.
Check your mix
Before the next long booking, run your reservation mix through the estimator — it does this division the regulation's way, then runs the hour tests and the loss gates on top. For the division on its own, there is an average stay calculator.
Common questions
How do you calculate average stay for a short-term rental?
Total guest nights divided by the number of reservations. It is not divided by 365, and not by the days the property was available. 180 nights across 50 reservations averages 3.6 nights.
Does passing the 7-day test make my losses non-passive?
No. It only removes the automatic passive label that attaches to rental activities. You must separately materially participate under one of the seven tests before a loss can offset other income.
Do long bookings ruin the average?
They can, faster than most owners expect. A handful of month-long winter stays mixed with weekend bookings can push the average past seven nights even when most reservations are short.