Screened 46 tracts down to 4 over six weekends, before spending a dollar

This is the boring kind of thing I actually want to read. Six weekends and no money spent and now you have four instead of forty six.

Did you find any tract that passed everything except the 2027 eligibility? I'd be curious whether you'd still build there.

Gate three is the one I'd argue with. One rent comp above 1,150 within a mile is a single data point, and in a distressed tract a single comp is often a one-off renovation someone overpaid to do. I'd want three comps and I'd want to know the lease dates.

Also your gates are all pass/fail, which means a tract that barely fails one dies while a tract that barely passes all four survives. That's fine for cutting 46 to 4 fast. It's not fine for ranking the 4, and ranking the 4 is your next problem.

bramble what did gate one actually look like in practice? I've got a tract that's currently designated and I'm fairly sure isn't on the new eligibility list, and I've been treating that as a soft negative rather than a hard gate. You made it a hard gate. Why?

On compass's question, the distinction that matters is what the designation status affects. A current designation governs whether an investment made now can qualify. Eligibility for the 2027 round tells you something different, which is how that census tract looks on whatever data screen Treasury applied for the new designation. Those are two separate facts and treating the second as a hard gate is a judgment about neighborhood trajectory rather than a tax conclusion.

The mechanism worth being precise about: the current set of designated zones sunsets at the end of 2026 and a newly designated set takes effect January 1, 2027, with the enhanced benefits under the One Big Beautiful Bill Act attaching to investments in the new zones. Whether a specific investment qualifies for which treatment depends on timing and structure that only your tax advisor can confirm against your facts in writing.

ledger's point about pass/fail gates is the operational one. A screen that cuts 46 to 4 is doing sourcing work, and sourcing work is allowed to be crude. Ranking survivors needs weighting, and the weight that usually decides these projects is not rent comps. It is cost to build and how long the jurisdiction takes to let you build. A tract with 1,250 rents and a nine month utility queue underperforms a tract with 1,150 rents and a four month queue on nearly every schedule I have seen modeled.

bramble, one gap in the four gates. Nothing in the screen tests whether an acquisition in that tract can plausibly clear the substantial improvement threshold at a cost that pencils. A tract full of recently renovated stock with high basis in the building is harder on that test than a tract of vacant shells.

Six weekends is the number I needed. I've been trying to work out what my first deal costs and I keep forgetting to price my own time before the deal even exists. Copying the four gates, swapping gate four because I've never dealt with any utility yet.