Yes, one long booking can do it, and the arithmetic is simple enough to check monthly. The short-term rental exception in Treas. Reg. §1.469-1T(e)(3)(ii)(A) applies when the average period of customer use is seven days or less, and that average is total days of customer use divided by the number of stays for the year. A quiet property with twelve short stays and one ninety-night winter lease averages nearly ten days and is a rental activity under the passive-loss rules, no matter how many hours the owner logged. A busy property with forty short stays can absorb the same booking without noticing. The only way to know which one you are is to keep the count.
The formula
The regulation (Treas. Reg. §1.469-1T(e)(3)(iii)) computes the average period of customer use by dividing the total days of customer use during the year by the number of periods of customer use. In practice, for a property booked by the night, owners and CPAs use nights as the unit:
Average stay = total nights booked ÷ number of stays, per activity, per tax year.
Three things about that formula are easy to miss. It is an average, so one outlier moves it. It is per year, so a great January does not offset a slow November; the whole year is one calculation. And it is per activity, which for most owners means per property, since grouping rentals into a single activity is an election with its own rules. The seven-day average stay guide covers the definition in more depth.
Why your minimum-night setting is irrelevant
Owners sometimes assume that a two-night minimum on the listing settles the question. It does not. The test looks at what guests actually did, not what the listing allowed. A minimum-night setting limits the shortest stay; it says nothing about the longest. The bookings that move the average are the long ones, and a minimum-night rule has no effect on those.
The same applies to the platform's own "average length of stay" statistic, if you have one. It may be computed over a different period, may exclude direct bookings, and may count in days rather than nights. Use it as a hint, not as the number.
Worked examples
Consider three properties over a tax year.
| Property | Short stays | Long stays | Total nights | Stays | Average |
|---|---|---|---|---|---|
| Busy beach condo | 40 stays × 4 nights | One 45-night off-season stay | 205 | 41 | 5.0 |
| Seasonal cabin | 12 stays × 3 nights | One 90-night winter lease | 126 | 13 | 9.7 |
| Mid-volume townhouse | 15 stays × 4 nights | Two 30-night relocation stays | 120 | 17 | 7.1 |
The condo is fine. The cabin is not, and the owner may not find out until the return is being prepared. The townhouse is the instructive one: it misses by a tenth of a night. Two thirty-day bookings that each looked sensible in the moment, taken together against a modest count of short stays, tip the year.
Notice that in every case the short stays are what protect the average. The more of them there are, the more room a long booking has. Volume is a cushion; low volume is exposure.
Monitor monthly, not in March
The failure mode is discovering the average in tax season, when the year is closed and nothing can be changed. The fix is to look at the running number every month, the same way you look at the hours total.
The running calculation needs only two figures per property: how many stays so far, and how many nights they added up to. Both belong to your booking platform's reservation report, which is the only place the list of stays is complete. A stay log kept by hand usually is not, and an incomplete list produces a confidently wrong average, which is why STR Tracker shows a reminder to check the figure at the source rather than computing it for you.
A useful habit is to note, alongside the current average, how long one more booking could run before the average crosses seven. That single number answers the question owners actually have in October: what does this inquiry for a three-week stay do to the year?
Options if you are drifting
This is a monitoring topic, not a booking-policy topic. Whether to accept a long stay is a business decision that involves revenue, wear, and your own circumstances, and no one should turn away a guest on the strength of a blog post. What the running number gives you is information in time to think, and a few things to think about:
- Talk to your CPA before, not after. If the average is heading over seven, a professional can tell you what the year looks like under the alternative rules, including the 30-day exception with significant personal services in §1.469-1T(e)(3)(ii)(B), and whether anything about your facts changes the analysis.
- Understand which year is affected. A long stay that straddles December and January raises its own counting question that practitioners handle differently. Ask.
- Keep logging hours regardless. If the activity turns out to be a rental activity for the year, the hours still matter for anyone pursuing real estate professional status, and a good log is never wasted.
The STR loophole guide explains how the seven-day test fits with material participation and why both have to hold in the same year.
Direct bookings and off-platform stays count too
Every period of customer use goes into the calculation, not just the ones a platform reported. Direct bookings from repeat guests, a friend-of-a-friend who paid by check, an insurance placement, a stay arranged through a local agency: all of them are stays with nights attached. Leaving them out understates the count and, if they were long, understates the average. A stay counter that only mirrors one platform is not a complete record.
The same completeness applies in the other direction. Personal use days are not customer use and do not belong in the stay count, though they matter for other rules.
Per property versus combined
The average is computed for the activity. If each property is its own activity, each gets its own average, and a long-stay property does not drag down a short-stay one. If properties have been grouped as a single activity, the calculation is combined. Grouping has consequences that reach beyond this test and is a decision to make with your CPA, in writing, not something to infer at year end from whichever calculation gives the better answer.
What to write down
For each property, for the year: every stay, its check-in and check-out dates, and the source of the booking. From that, stays and nights fall out, and the average follows. Update it when bookings land, not when the return is due. If you are not tracking it yet, you can start a free trial and enter this year's totals in a few minutes, then watch the number instead of guessing at it.