I priced 12 lots off asking prices in a state that doesn't publish sale prices

Your own conservative number, $54,000, already showed a $98,000 loss before commissions. That's the sentence I'd tape to the wall. The comp problem made it worse, sure, but the model told you no at your own low case and you overrode it. I've done a version of that with a different asset class and it's the most expensive habit I have.

Working your numbers per acre so the shape is visible. You paid $7,900 an acre bulk. Your improvements and soft costs added $3,750 an acre, so you're at $11,650 an acre of cost. Twelve lots averaging 5.3 acres means your cost per lot is about $62,000. You sold at $46,900. You were underwater per lot at the moment you closed on the land, regardless of comps.

So I don't think comps were the primary failure. The improvement cost per acre was too high for the retail price per acre this market supports. $186,000 of road for 12 lots is $15,500 a lot, and on $47,000 lots that's a third of gross. In markets with $47,000 lots, the deals that work are usually splits off existing frontage where the road cost is near zero, or bulk land bought at half of what you paid.

The absorption assumption is the second problem and it's just as costly. Eight lots in 19 months is one per 2.4 months, which is not terrible. Your interest was $71,400 across the hold, about $3,760 a month, and that's the meter running against every month of marketing time. Any deal where carry is that heavy needs either a much bigger spread or a shorter path to first closing.

Coming from trades, the $186,000 road number is the one I'd want broken down. 900 feet with 12 lots hanging off it. Was that all base and grading, or were there utilities, drainage structures, and a stormwater pond in there? I ask because the drainage part is where I've seen numbers double, and if the county made you build a detention basin, that's not really a road cost, it's an approval cost you couldn't avoid.

On the non-disclosure state point, there's a structural workaround people forget. Deeds are still recorded even where prices aren't, and in a lot of counties the recorded transfer tax stamp or a declaration of value filing lets you back into a price, and in some states the assessor's ratio study data is available on request. Whether any of that is available depends on the state and sometimes on the county clerk's mood. Worth an afternoon at the recorder's office before you spend $500,000.

And get the agents' opinion of value in writing with the closings listed, like you said. A written opinion also gives your lender something to look at, which matters when you go back for the next facility.

Two separate assumptions failed here and they need separating, because the fixes are different.

The retail price assumption failed because asking prices measure what sellers want. In states where sale prices aren't public, and roughly a dozen states don't publish them, the substitutes are a licensed appraiser's lot valuation, brokers' written opinions of value backed by transactions they closed, builders who buy lots in that submarket, and where available recorded transfer declarations. Practice varies by state and by county on what's obtainable, so confirm locally. Any of these costs a small fraction of a $500,000 acquisition.

The cost assumption failed independently. At $11,650 per acre of total cost against $8,850 per acre of realized revenue, the deal never had a spread. This is the arithmetic that decides subdivide deals: bulk purchase price per acre, plus improvement and soft cost per acre, plus carry, against realized retail per acre net of commission. A minor split off existing frontage carries almost no improvement cost, which is why low-priced markets tend to support those and not new-road subdivisions. When retail lots are under $50,000, a $15,500 per lot road allocation consumes the spread by itself.

On absorption, interest at roughly $3,760 a month means every month of unsold inventory costs about 8% of one lot's price. Underwriting absorption at half the hoped pace, as you said, is the right correction, and it should be run as a sensitivity before purchase rather than a comfort statement.

The four remaining lots are now a separate decision from the eight sold. Their relevant question is the price that clears them against continued carry, not recovering the loss on the first eight.

The line that stuck with me is that you told yourself commissions were negotiable and you'd sell three yourself. Every thin deal I've been shown has one of those, a cost that becomes flexible in the model right when the model needs it to be. Did you end up selling any without an agent?

I've got two lots on the wet end of a split sitting right now and the answer for me has been to stop pretending they're lots. One's going to a neighbor as an addition to his acreage at bulk pricing, which hurts, and it's still cheaper than three more years of interest. Your four might be the same conversation. Who owns adjoining you?

Scope creep on renovations at least announces itself, you open a wall and see it. Land scope creep arrives as a letter from an engineer. What I recognize in your post is the moment where the number went from $54,000 to $58,000 because you needed it to. I've moved a resale number by $15,000 for exactly that reason and paid for it later.