My house-hack calculator told me to rent. Barely.

In March I was two weeks from signing on a townhouse in Halethorpe, Maryland. UMBC-adjacent block, walkout basement with its own kitchenette and full bath, two bedrooms upstairs sharing a hall bath.

Pre-approval in hand. Buyer's agent referred to me by the mortgage officer, with a track record. I'd told her the bar I cared about: a property that would beat the S&P 500 over a 10-year hold, owner-occupied, with rentable space inside the same address while I lived there. I was open to either house-hack path — a 2–4 unit FHA 3.5%-down play, or a single-family with a rented basement on a first-time-homebuyer 3%-down conventional — whichever cleared the bar. She brought me a townhouse, so the path was the 3%-down conventional.

She had one. $299k. The basement bathroom needed a full redo. About $20k. Another $5k to furnish the basement and replace the basement carpet. And the upstairs would need separate washer-dryer hookups installed before it could be rented as its own unit. Tolerable on a 3%-down loan that left me roughly $52k of cash committed at close. That's down payment, closing costs, the basement work, and a 3-month emergency fund the model parks in cash before any rental upside compounds. The kitchen was the right kind of bad. I was emotionally close.

At the open house she told me I could easily rent the basement for $1,300. I checked comps that night. The most expensive basement rental in the neighborhood was $1,100 — I'd have been the priciest by $200, in a price tier that's already at the top of what tenants pay for finished basements in that ZIP. She wasn't being dishonest. She was rounding the rent up by about 18%, the way every real-estate pitch does. And 18% on the rental side is the difference between "this beats the index" and "this is a coin flip."

I asked for the underwriting. What's the cap rate. What does cash flow look like once I move out and rent the whole house. What's the breakeven if the upstairs sits vacant for three months. She kept saying she'd send it.

She didn't. She sent the contract first.

---

I took a Saturday and rebuilt the math from scratch.

Not because I thought she was a scammer. She probably runs the same playbook on every buyer, where "the numbers" are a comp sheet from Zillow and a vibe check. I rebuilt it because a benchmark-vs-alternative comparison is harder than it looks if you're being honest, and I didn't trust anyone else to do it for me.

I went looking for a calculator that would let me compare a house-hack against the index honestly, and the public calculators sort into two camps. Among the dozen-plus that market themselves specifically for house hacking — BiggerPockets, SparkRental, ADPI, RFP Homes, Everything House Hacking — none tax the index alternative, none surface a breakeven where the verdict flips, and none disclose §469 / §1250 / §121 as material unmodeled risks. The general rent-vs-buy calculators that do tax the alternative (BuyVsRent.org, the NYT calculator, TrueHousingCost) are owner-occupied only. They don't model rental income from the rented unit, which is the entire point.

So I built mine into the gap.

I credit rental-portion depreciation (27.5-year straight-line on the non-land basis of the rented portion) and rental-portion mortgage interest at your marginal tax rate, opt-in. The "Tax Savings" line is an upper bound. Here's what I don't model: tax owed on positive net rental income, §469 phaseout above $150k MAGI (which suspends most of the published savings until exit), §1250 recapture at sale (a one-time ~25% tax on accumulated depreciation), §121 exclusion that only covers the residence portion of a mixed-use property, state income tax, and S&P long-term capital gains on realization. The realistic after-tax benefit captured on the property side is typically 40–70% of what the headline shows.

Everything I don't model would only widen the S&P advantage further. They all bias against the leveraged real-estate position. The leaning verdict you're about to read is the conservative case for the property, not the punitive one. None of this was in the comps the agent didn't send.

---

The Halethorpe townhouse, run cautiously — 12% maintenance + vacancy, 22% federal bracket, 3% appreciation, 2-year owner-occupy then 8 years fully rented, first-time-homebuyer conventional at 3% down and 6.875% — came back leaning S&P over a 10-year hold.

Both paths got rich in absolute terms. 17× on the property side, 19× on the index. Leverage on one, a decade of compounding on the other. Neither path is a disaster. The S&P just wins by enough.

Not the dramatic miss you'd hope for to make the decision easy. The verdict band the calculator outputs in that range is Leaning — real advantage, not overwhelming. The kind of close call where a competent underwriter says "this is borderline, what's your tolerance for being wrong."

That's the story most house-hack content skips. The dramatic miss makes for a better tweet. The close call is more common, and the close call is where the methodology earns its keep.

The agent kept telling me this property had everything I'd asked for. Perfect for your goals. But Halethorpe didn't have a single funded growth catalyst. No MARC station expansion, no metro extension, no Baltimore County capital plan that even mentioned the area. Appreciation would track Maryland averages at best. The math was already showing leaning S&P, and there was no underwritten upside that could plausibly tip it back toward the property over a decade.

The basement wasn't comfortable. The floor didn't feel level, and an exposed HVAC duct in the bathroom forced me to lower my head to wash my hands. Felt like a garage someone had legally permitted as a unit.

None of that stopped me from being two weeks out. I was hypnotized by the classic realtor sales-y stuff: the urgency, the perfect for your goals. The calculator was a bucket of water.

So I asked the question the verdict pushed me to: was this really worth the hassle of being a landlord? Fielding the 2 AM kitchenette calls and eating the §1250 recapture I knew was coming at exit. All for a margin the index would erode anyway.

I walked.

The honest read is that the math came back as marginal-but-leaning, and a stack of soft signals — the missing catalyst, the inflated rent assumption, the agent's enthusiasm where the numbers didn't earn it — decided. I'm not pretending the calculator made the call alone. It gave me a defensible bar to walk away from, and that bar held when the soft signals went bad.

---

The townhouse is still listed. 61 days on market as I'm writing this. The market is more or less voting along with my calculator. At the asking price, this isn't a great property. Eventually someone will run the math differently, or accept a lower margin than I would, and the listing will close. The model is an underwriting standard. It tells you whether the deal clears the bar at the assumptions you can defend, and what would have to be true for it to clear at the assumptions you can't.

That's the second-order question, and most public real-estate calculators skip it entirely. You enter a price, a rent, a rate, an appreciation. They give you the answer at one set of inputs. They don't tell you which inputs the answer is most sensitive to, and they don't show the breakeven where the verdict flips. §469, §1250, and §121 aren't modeled at all, because those tax events fire at exit, and most calculators are sold by people who don't want you thinking about exit.

If you find an input where the verdict reverses, or a tax event I should model instead of just disclose, email danny@therenthacker.com or find me in the thread. That's where this gets better.