To calculate ARV, find recently sold homes near the property that were renovated to the standard you plan to hit, divide each sale price by its finished square footage, take the median of those figures, and multiply it by your property's square footage. A 1,600 ft² house in an area where renovated comps sell for a median of $241/ft² has an ARV of about $385,600. ARV is a comp-based estimate, not a formula you can derive from the property alone.
Written down it looks like a one-line calculation, and that is misleading. The second term you read off the tax record. The first term is a judgment call about which sales in a neighborhood describe the house you are going to hand back to the market, and it is where every ARV argument actually happens. A generous set of comps and a conservative set of comps on the same street routinely differ by 15%, which on a $385,000 house is $58,000 — more than most flip margins.
So treat the steps below as a procedure for defending the first term, not for doing the multiplication.
Decide the boundary first — typically a half-mile radius, tightened to a quarter mile in dense urban grids and loosened to a mile in rural areas, and never crossing a school-district or municipal line. Deciding after you have seen the sales is how a $460,000 sale two neighborhoods over ends up "comparable."
Active listings are asking prices, which are opinions. Pending sales have a number but not a confirmed one. Only recorded, closed sales tell you what someone actually paid. Six months back is the standard window; go to twelve only if you must, and expect to defend it.
This is the step that separates ARV from ordinary comping. You are not looking for houses like the one you are buying — you are looking for houses like the one you will be selling. If your scope is new kitchen, two new baths, floors and paint, your comps are renovated sales, not tired ones.
Divide, list them out, and take the middle value. Keep the list visible; you will need the spread later.
The median times your square footage is your ARV. The lowest and highest comp times your square footage is the range that ARV honestly sits in — and quoting the range alongside the point estimate is the single easiest way to stop yourself over-trusting it.
Subject property: 1,600 ft², 3 bed / 2 bath, single-story, dated but structurally sound. Planned scope is a full cosmetic renovation — kitchen, both baths, flooring, paint, and mechanical updates — budgeted at $55,000. Here is the comp set after the filters above:
| Comparable sale | Sold price | Sq ft | $/ft² | Distance | Closed |
|---|---|---|---|---|---|
| 118 Ashwood Dr | $412,000 | 1,710 | $241 | 0.3 mi | 6 weeks ago |
| 42 Belmont Ave | $359,000 | 1,520 | $236 | 0.4 mi | 3 months ago |
| 907 Crestline Rd | $437,000 | 1,730 | $253 | 0.5 mi | 2 months ago |
| 15 Danbury Ln | $351,000 | 1,610 | $218 | 0.2 mi | 5 months ago |
| 630 Elmhurst St | $404,000 | 1,660 | $243 | 0.4 mi | 7 weeks ago |
Sorted, the price-per-foot figures run $218, $236, $241, $243, $253. The median is $241 and the mean is $238 — close together, which is the reassuring case: no single sale is dragging the set.
And the honest range, from the lowest and highest comp: $348,800 to $404,800. That $56,000 spread is not a failure of the method — it is what the neighborhood's own sales disagree by, and pretending otherwise is how people talk themselves into deals.
Faro pulls the comparable sales for any address, computes the value from both listed and recorded sales, and shows you the confidence band around it. Enter the address to estimate ARV and analyze the deal.
Price per square foot already handles size, roughly. What it does not handle is everything else, and on a small set those differences matter. Adjust the comp toward the subject — if the comp has something yours won't, subtract from the comp's price; if yours will have something the comp lacked, add.
| Difference | Typical adjustment | Note |
|---|---|---|
| Extra full bathroom | $8,000 – $15,000 | Bigger in smaller houses, where a second bath changes who can buy it |
| Garage vs. no garage | $10,000 – $20,000 | Strongly climate-dependent; near zero in some markets |
| Finished basement | Exclude from ft², adjust separately | Counting it in square footage is the single most common comping error |
| Lot materially larger | Local $/acre or $/ft² of land | Only worth adjusting past roughly a 25% difference |
| Backs onto a highway or commercial | −5% to −12% | Look for a matched pair on the same street to size it |
| Sold 9–12 months ago | Local price trend × months | Prefer replacing the comp to adjusting it this hard |
Two rules keep this honest. Adjust as little as possible — a comp needing three adjustments is telling you it is not a comp. And never adjust in only one direction; a set where every adjustment happens to raise the ARV is a set that was reverse-engineered from the answer you wanted.
Before you rely on the number, look at what being wrong does. Same house, same $55,000 of repairs, the 70% rule applied to each ARV — this is the maximum you could pay and still be following the rule:
| If ARV is really… | 70% of ARV | Max offer after repairs | vs. base case |
|---|---|---|---|
| $366,300 (−5%) | $256,410 | $201,410 | −$13,500 |
| $385,600 (base) | $269,920 | $214,920 | — |
| $404,900 (+5%) | $283,430 | $228,430 | +$13,500 |
A 5% error in ARV moves what you can pay by $13,500, and it moves your profit by considerably more than that because the error lands entirely on the sale side. This is why the range matters as much as the point estimate, and why the honest response to a wide comp spread is a lower offer rather than a rounder number.
With an ARV and a repair budget, the maximum allowable offer is one more step. The 70% rule calculator shows the offer and the margin it reserves for holding, selling, and profit.
Open the 70% rule calculatorARV = comparable price per square foot × your property's square footage. The formula is trivial; all of the work is in the first term. Two people can apply it to the same house and land $60,000 apart purely on which sales they accepted as comparable.
Three is the working minimum and five to six is comfortable. Below three you have no median worth the name — one unusual sale moves the answer by tens of thousands. Above about eight you are usually reaching outside the neighborhood or back in time to fill the set, which costs you more accuracy than the extra data point buys.
You can estimate ARV without ever going inside, because ARV describes the finished house, not the current one. What you cannot estimate from the outside is the repair number — and since the offer is ARV minus repairs, a sight-unseen ARV paired with a guessed repair budget is only half an analysis. See ARV for wholesaling for how to work when that is all you have.
No, though a good ARV predicts one. An appraiser values the house in its current condition unless they have been given a scope of work and asked for a subject-to-completion value. A refinance lender on a BRRRR will order exactly that kind of appraisal — which is why an ARV built on comps an appraiser would also accept is worth more than an optimistic one.
Median. One flip that sold to an emotional buyer, or one comp with a finished basement counted in its square footage, drags an average somewhere no ordinary sale will follow. If your median and your average are far apart, that is a signal in itself: it means the set contains something that does not belong.
Where the sold data actually lives, and the six filters that separate a comp from a nearby house.
Two different numbers on the same house — where the gap comes from and who uses which.
The maximum allowable offer formula, both versions, and how to pick the percentage honestly.
Matching your scope to your comps, and the sensitivity table that shows what a small ARV miss really costs.
Enter an address and Faro pulls the comparable sales itself, then shows the value, the rent, and the price you'd have to buy at.
Estimates for analysis and educational use only — not financial, investment, tax, or legal advice. Verify every number independently before making a purchase decision.