The Cheapest Square Footage in Richmond Sat for 70 Days. We Took It Apart.

One filter, 102 houses, one survivor

Most Richmond market coverage tells you the median sale price moved. That number is one figure, reported a month late, averaged across a metro that spans $161-per-foot in South Richmond and $493-per-foot in the Museum District. You cannot buy the median. You cannot negotiate against it.

So we asked a narrower question, one that has an actionable answer: which houses are still available after the market has had a real look at them? Richmond homes that go under contract mostly do it fast. In our active-inventory snapshot the median listing had been up for 23 days, and a quarter of the inventory had been up for 3 days or fewer. A house that is still sitting after ten weeks has been seen, toured, and passed over — repeatedly, by people who were actively shopping in exactly its price band.

That is not automatically a bad house. It is reliably an informative one. Time on market is the single cheapest piece of negotiating leverage available to a buyer, and it is printed on the listing for free.

The filter is deliberately unforgiving. A listing qualifies only if it has been live for more than eight days — long enough to have cleared its first weekend of showings without going under contract — and only if it is genuinely purchasable. Pending, contingent, under-contract-taking-backups, coming soon, recently sold, auction, foreclosure: all excluded, because a house you cannot buy is not a recommendation, it is a tease. Of 102 scored active listings, 65 cleared the bar.

Then we ranked the survivors on the gap between what they were asking and what the surrounding data implied they were worth, with time on market and any published price cut as amplifiers. One house came out well ahead of the rest.

10119 Cravensford Ter, Midlothian 23112

The facts, stated plainly. Four bedrooms, four baths, 4,285 finished square feet on 0.36 acres, listed at $629,000 in the 23112 ZIP of Midlothian, Chesterfield County. That works out to $147 per finished foot. Zillow's Zestimate at capture was $626,600 — within half a percent of the asking price.

For scale: the median active listing in 23112 was 3,413 square feet asking $722,475. This house is 26% larger for 13% less money. That is not a rounding difference. That is a different pricing regime, sitting inside the same ZIP code, on the same market, at the same time.

Four methods, four answers, a $320,000 spread

Here is where it gets genuinely interesting. Run this house through four ordinary valuation approaches and they do not converge — they fan out across a range wider than the price of a starter home.

Three of those four methods say this house is dramatically underpriced. One says it is priced exactly right. The market — which is the only method that clears cash — said nothing for seventy days.

Note which method disagrees with the others, because it matters. The Zestimate is the only one of the four that is not purely a size-and-location extrapolation. It ingests condition signals, photos, tax records, and — critically — the listing's own price history and market response. When an automated valuation lands within half a percent of the ask on a house that every $/sqft comparison calls a steal, the AVM is quietly telling you it agrees with the sellers, not with the models.

That is not proof the models are wrong. AVMs anchor hard on list price and are notoriously weak on unusual properties, which is exactly what this is. But it is a real signal, and it points the same direction the seventy days do.

What $147 a foot looks like against its neighbors

Here is every active 23112 listing that was asking under $175 per square foot at capture, plus the ZIP's median for context. This is the value tier of the neighborhood — and the subject sits at the bottom of it.

Two things jump out of that table, and the second is more important than the first.

First: the subject is the cheapest per foot in its own ZIP, and the cheapest per foot of any home over 4,000 square feet anywhere in the metro slice — 36 listings, and this one is at the bottom.

Second, and this is the part worth sitting with: the cheap-per-foot tier of this ZIP is bimodal on time. 11007 Wooferton at $166/sqft had been up two days. 6024 Watch Harbour at $173/sqft had been up two days. 2719 Mill Flume at $151/sqft — barely more expensive per foot than the subject, and larger — had been up fifteen. Meanwhile the two homes that had sat longest in this tier are the subject at 70 days and 13805 Beechwood Point at 112.

Low $/sqft in 23112 is not, by itself, what makes a house sit. Homes at nearly the same price per foot were moving. Something specific to these two houses is doing the work.

What seventy days actually means

Days on market is the most misread number on a listing sheet, so it is worth being precise about what we can and cannot say from this data.

At 70 days the subject sits just past the 75th percentile: 24 of the 100 active listings with a recorded day count had been up longer, 76 had been up less. It is not an outlier. It is not a derelict that has been rotting on the MLS for a year. It is squarely in the slow quartile — the zone where sellers have stopped expecting a bidding war and have started doing arithmetic about carrying costs.

Why big, cheap houses are the most mispriced category in any market

There is a structural reason a 4,285-square-foot house at $629,000 confuses valuation models, and understanding it is worth more than this one listing.

Price per square foot is not linear. It falls as homes get larger, because the expensive parts of a house — kitchen, baths, mechanicals, the roof over the first thousand feet — do not scale with floor area. The marginal square foot of a big house is a bedroom or a bonus room, and bedrooms are cheap to build. Our own metro data shows it directly: across the 41 homes over 4,000 square feet in the August sold set, the median was $246 per foot on a median price of $1.2 million, while the metro-wide median across all 171 sales was $227 per foot on a $750,000 median.

So a model that prices a 4,285-square-foot home by multiplying square footage against a ZIP-wide median $/sqft — a median dominated by 3,400-foot houses — will systematically overshoot. That is a large part of the $249,802 gap between the hedonic model's $878,802 and the $629,000 ask. It is not the model finding hidden treasure. It is the model doing arithmetic on a house type it has few examples of.

The honest read is narrower and more useful: compare the subject against homes of its own size, not its ZIP's median size. Against the 36 active homes over 4,000 square feet in the metro, the median was $190 per foot, which would imply $814,150. The subject asks $629,000. That is still a 23% discount — smaller than the headline 51%, and far more defensible.

Five things that make a house sit at a price like this

When homes at nearly the same price per foot are selling in two weeks and this one has not moved in ten, the explanation is almost always one of five things. Every one is checkable before you write an offer, and every one is a different negotiation.

• •Condition. The most common answer by a wide margin. A 4,000-plus-square-foot home with original mechanicals, an aging roof, or dated kitchens and baths carries a renovation bill that scales with its size. $80,000 of deferred maintenance on a $629,000 house is 13% of the purchase price, and buyers price it in at retail, not at contractor cost.
• •Layout and functional obsolescence. Square footage is not fungible. Finished basement space, low-ceilinged bonus rooms, a fourth bedroom that is only reachable through a third, or a floor plan that puts every bedroom upstairs from a primary on the main — these count fully in the $/sqft calculation and count for very little to a buyer.
• •Location within the ZIP. 23112 is large and not uniform. Backing to a road, sitting under transmission lines, an unusual lot orientation, or a stretch of the neighborhood with a different builder and vintage will not show up in a ZIP-level model at all.
• •A condition the listing does not advertise. Septic rather than public sewer, a well, a flood-zone designation, an easement, an HOA in litigation, a special assessment, or a pool the next owner does not want. Any one of these can put a house outside a large fraction of buyers' financing or appetite.
• •Nothing at all — it is genuinely mispriced. This does happen, most often with an out-of-area listing agent, an estate sale, a relocation, or a seller who set the price against a comp that was never comparable. It is the least likely explanation and the only one that makes the discount free.

Notice that four of the five are discoverable in a single afternoon: a walkthrough, the county's assessment and permit records, the plat, and the seller's disclosure. The fifth reveals itself when nothing else does.

What it costs to own, at four different offers

Carry cost is where a negotiation stops being about pride. Below is the monthly obligation at four price points, assuming 20% down, the Freddie Mac 30-year fixed average of 6.67% as of August 13, 2026, Chesterfield County's real estate rate of $0.89 per $100 of assessed value, and homeowners insurance at 0.45% of value annually. No HOA is included — 23112 has many neighborhoods that carry one, and it is a question to ask before you model anything.

The spread between asking and a $550,000 offer is $494 a month — about $5,900 a year, or $178,000 across a thirty-year term before any refinance. On the other side of the table, if the seller is still carrying a mortgage on this house, seventy days of waiting has plausibly cost them somewhere in the region of $9,000. That number is the reason a seasoned listing agent gets nervous in week eight, and it is why time on market converts to leverage.

It also frames the renovation math. If a walkthrough turns up $80,000 of work, the question is not whether the house is worth $629,000. It is whether $629,000 plus $80,000 plus the months of disruption beats the $814,150 that a comparably sized, move-in-ready home in this metro would cost. Often it does. That is the actual trade, and it is a trade about the buyer's tolerance for a project, not about whether the listing is a bargain.

Schools, and a caution about school claims

23112 sits in Chesterfield County Public Schools. The elementary, middle, and high schools serving this ZIP in our dataset include Woolridge, Spring Run, Swift Creek, Clover Hill and Alberta Smith at elementary; Swift Creek and Bailey Bridge at middle; and Cosby, Clover Hill and Manchester at high school. Free-and-reduced-lunch rates across that set run from 8.5% at Woolridge to 82.8% at Thelma Crenshaw — an enormous spread inside a single ZIP code.

How we would actually approach it

Diligence first, and in this order. The answers determine the offer, not the other way round.

• •Pull the county assessment and permit history. Chesterfield publishes both. The assessment tells you what the county thinks it is worth and, when it lands far below the ask, gives you a documented third-party number to negotiate against. The permit history tells you whether the roof, HVAC and water heater have ever been replaced.
• •Establish what the 4,285 square feet actually is. How much is above grade, how much is finished basement, how much is a bonus room over the garage. Ask for the appraiser's or assessor's breakdown rather than the marketing figure. This single question explains most of the $/sqft anomaly in most cases like this.
• •Walk it with a contractor before you write, not after. On a house this size an inspection contingency negotiated after the fact is a weaker position than a number you brought with you.
• •Ask the listing agent one direct question: what feedback have showings given? After seventy days there is feedback. Agents will frequently tell you, and it is the cheapest information in the transaction.
• •Check sewer, water, flood designation and HOA status. Septic, well, a flood-zone overlay or a special assessment each materially change the number and each is a public record.

On price: the defensible anchor is not the ZIP median and not the hedonic model. It is the $190-per-foot metro median for homes of this size, which implies $814,150 for a comparable house in ordinary condition. The gap between that and the $629,000 ask is the market's estimate of what is wrong with this one. If the diligence turns up $80,000 to $100,000 of work, the ask is roughly fair and the negotiation is about closing costs and repair credits. If it turns up $200,000 of work, walk. If it turns up almost nothing, an offer near asking is defensible and you should move quickly, because a house that survives that diligence will not survive it twice.

The general lesson, which outlives this listing

Whatever happened to this particular house, the method transfers, and it is the reason we are going to do this every week.

Days on market is public, free, and almost universally under-used. In a metro where the median active listing had been up 23 days, everything past about eight weeks is a seller who has already revised their expectations at least once, whether or not they have revised the price. Price per square foot is the most useful comparison available to an ordinary buyer and the easiest to misuse — it must be run against homes of comparable size, not against a ZIP-wide median, or it will manufacture bargains that do not exist. And when an automated valuation and a regression model disagree by $250,000 on the same address, the disagreement itself is the finding. It tells you precisely where to point your diligence.

The houses that sit are not the leftovers. They are the only listings where a buyer has time to think, information to work with, and a seller who has run out of both.

Sources and methodology

• •Subject listing, active-inventory metrics, DOM distribution, 23112 comparison set — src/data/for-sale-scored.json, hedonic scoring run v3.1.0 generated 2026-03-12, 102 scored active listings. Fields used: listPrice, sqft, beds, baths, lotAcres, daysOnMarket, ppsf, zestimate, predicted, deltaPct, verdict.
• •Recent-sold medians and the large-home $/sqft comparison — src/data/rva-sf-4bd-2250sf-sold-30d-2026-08.json, 171 Richmond-metro single-family sales scraped 2026-08-16 over a 28-day window (median $750,000, median $227/sqft), and src/data/sf-rva-area-sold-30d-2026-07.json, 558 sales scraped 2026-07-12. The 30-day maximum on daysOnZillow in both files is a search-window artifact, discussed above; no DOM claim in this piece is drawn from either.
• •Chesterfield County real estate tax rate, $0.89 per $100 assessed — src/data/mortgage/locality-tax-rates-2026.json, last verified 2026-05-04 against chesterfield.gov.
• •Mortgage rate, 6.67% on the 30-year fixed — Freddie Mac Primary Mortgage Market Survey, week of August 13, 2026 (6.69% the prior week; 6.58% a year earlier).
• •School enrollment, student-teacher ratios and free-and-reduced-lunch rates — src/data/richmond-schools.json, NCES-derived, Chesterfield County schools with a 23112 service address.
• •Selection logic — src/lib/featured-property.ts and src/lib/listing-status.ts. Eligibility and ranking were not applied by hand for this article; the property was selected by running the same code the weekly digest runs.
• •Estimates are labelled as estimates. Carry-cost figures assume 20% down and 0.45% annual insurance and exclude HOA dues, PMI and closing costs. The renovation figures are illustrative thresholds for framing a negotiation, not an appraisal of this house.