To calculate ARV (after-repair value), pull at least three closed sales of similar houses in the same market area, adjust each one up or down to match your property, then reconcile the adjusted figures into a single number. That number is roughly what your house is worth once the work is done. The order matters more than anything else on this page. Averaging the comps' raw price per square foot and adjusting afterward is the sequence most beginners use, and it runs high. Here is the right order, with the arithmetic shown.
Educational information only, not investment, legal, or appraisal advice. An ARV you calculate yourself is an estimate, not an appraisal, and nothing here is a promise that any number produces a profit. Confirm values with a local agent or appraiser before you commit money.
What ARV actually is
ARV = After-Repair Value = the market value of a property after it has been renovated to a normal, sellable condition. It answers one question. "If I fixed this place up, what would it sell for?" You use it to work backward to a safe offer, see what to offer on a house.
The clearest public definition of the idea sits inside FHA's rehab-loan program, which calls the number the After Improved Value and defines it as
After Improved Value refers to the value as determined by the Appraiser based on a hypothetical condition that the repairs or alterations have been completed.
That is FHA program text, from HUD Handbook 4000.1 (Update 18), II.D.11.c.ii, p. 862. It governs FHA 203(k) rehab mortgages, so it is not a universal legal definition of ARV. But two things follow from that one sentence and they hold wherever the number is used.
ARV is hypothetical. It describes a house that does not exist yet. Every figure in it is a claim about the future, which is why the honest version of an ARV is a range with a reason attached.
ARV is pinned to a written scope of work. The same FHA section requires the appraiser to review the consultant's Work Write-Up or the contractor's proposal, and to complete the appraisal based on the final write-up. Change the scope and the ARV has to change with it. An ARV that stays the same while the repair list shrinks is not an ARV, it is a wish. That is also why the repair estimate and the ARV have to be built together, not separately.
One honest note on sourcing. Fannie Mae's and Freddie Mac's appraisal sections do not mention ARV at all. What they give you is the machinery any credible ARV rests on, which is how comparable sales get chosen and adjusted. That is what the rest of this page uses them for.
The comp-based method (in the right order)
Step 1. Pull at least three closed sales
Fannie Mae's Selling Guide is the best public statement of how comps get chosen, because appraisers working on loans sold to Fannie Mae have to follow it. None of it is law for an investor. You can still run the same tests.
- At least three, and they have to have closed. "A minimum of three closed comparables must be reported in the sales comparison approach" (Fannie Mae Selling Guide B4-1.3-08). A pending sale or an active listing is supporting data only. "Contract offerings and current listings can be used as supporting data, if appropriate," says the same section, which is not the same thing as counting as one of your three.
- Twelve months, and it is a "should", not a must. "Comparable sales that have closed within the last 12 months should be used in the appraisal" (same source). Older sales are allowed where market conditions limit the recent ones, with an explanation. Fannie's section contains no six-month preference. Six months belongs to two other programs. Freddie Mac requires commentary on any comp more than six months old (Freddie Mac Seller/Servicer Guide 5605.6(g)), and FHA prefers sales within the last six months on FHA-insured loans, with at least three that settled inside 12 months (HUD Handbook 4000.1, II.D.4(3)(c)). Name the program or do not use the number.
- There is no mileage radius. Neither guide sets one. Fannie requires the distance to each comp to be stated honestly and "measured using a straight line between the properties" (same source), and allows rural comps a considerable distance away when they produce credible results and are explained. State the distance, do not hide behind a radius you invented.
- One market area, and you do not move the line to suit the answer. "the appraiser must not expand the neighborhood boundaries just to encompass the comparables selected" (same source). Comps from a competing market area are allowed only when the difference is disclosed and addressed.
- Match on what the house is. Comps should share physical and legal characteristics with the subject. "These characteristics include, but are not limited to, site, room count, finished area, style, and condition" (same source). They need not be identical, they need to compete for the same buyer. FHA states the corollary more bluntly than either GSE, that comps "should be selected based on similar locational and physical characteristics, not sales price" (HUD Handbook 4000.1, II.D.4(3)(e)(i)).
What about the tighter screens, the ones in our own coursework? Ours are the last 3 to 6 months, about half a mile, and finished square footage within about 10 percent, where those sales exist. Those are our screens, not anybody's rule, and no primary source sets them. They are usually easier to defend, because a comp that needs fewer adjustments is a better comp, but that is a craft argument and not a citation. Never present them as the rule, and be ready to widen them when the tight screen returns one sale instead of three.
Both guides get revised. Freddie's section 5605.6 has a new version that becomes mandatory on November 2, 2026 under the new appraisal dataset standard, so check the live text before you quote either one.
Step 2. Adjust each comp to your property
This is the step the whole method turns on. Every comp differs from your house, so you correct each comp's sold price for those differences before you combine anything. Add to a comp that is inferior to your subject. Subtract from a comp that is superior. What you are left with is what that comp would have sold for if it had been your house.
- There is no percentage cap on adjustments. "Fannie Mae does not have specific limitations or guidelines associated with net or gross adjustments" (Fannie Mae Selling Guide B4-1.3-09). Freddie says the same thing in its own words, "Freddie Mac does not have limitations on gross or net adjustment percentages" (Seller/Servicer Guide 5605.6(a)). The 15% net and 25% gross limits you will hear quoted as gospel are folklore, disclaimed in writing by both.
- The dollar amount comes from the market, not from a rule of thumb. Fannie's adjustments section calls a $20 per square foot rule-of-thumb adjustment inappropriate where market analysis supports $100 per square foot. You get the rate by finding pairs of sales that differ in one thing and reading what the market actually paid for that difference. And where the adjustments get large enough to suggest the property may not conform to its market, Fannie hands the question to the underwriter, which tells you the size of an adjustment is itself information about the comp.
- A time adjustment needs evidence, and so does skipping one. "Time adjustments, or the lack thereof, must be supported by evidence" (same source). Evidence means a home price index, paired sales, or statistical analysis. A feeling about where the market is going is not evidence in either direction.
Step 3. Reconcile the adjusted figures, do not just average them
Reconciling is not averaging. An average treats a comp that needed one small correction and a comp that needed four large ones as equally good evidence, and they are not. Fannie's own guidance makes the same point about recency, its example prefers a nine-month-old sale that needs one time adjustment over a one-month-old sale that needs several adjustments.
So weight the comps that needed the least correction, and read the spread. Three adjusted figures that land within a few thousand dollars of each other are a credible ARV. Three that land $40,000 apart are telling you the comps are weak, and the honest response to that is a wider margin, not a more confident number.
A worked example
Your subject is a 1,550 sq ft house, 3 bedrooms, 2 baths, no garage. You plan to renovate it to the finish level this market area sells at. Three closed sales, all in the same market area, all inside twelve months.
| Comp | Sold price | Living area | How it differs from your subject | Closed |
|---|---|---|---|---|
| A | $255,000 | 1,500 sq ft | 50 sq ft smaller, otherwise a match | 2 months ago |
| B | $276,000 | 1,600 sq ft | 50 sq ft larger, has an attached garage | 4 months ago |
| C | $267,000 | 1,500 sq ft | 50 sq ft smaller, has an extra half bath | 9 months ago |
Paired sales in this market area supported three adjustment rates. $60 per square foot of living-area difference, $12,000 for an attached garage, and $9,000 for a half bath. Those are this example's own figures, derived from its own market. Yours will be different, and Fannie's section is explicit that a rule-of-thumb rate is the wrong way to get them. Note also that $60 is nowhere near the roughly $170 per square foot these houses sold for. The marginal value of extra floor area is not the average price per square foot, and that single fact is why the two methods below disagree.
Now adjust each comp to the subject.
Comp A is 50 sq ft smaller, so it is inferior on size. Add.
$255,000 + (50 × $60) = $258,000
Comp B is 50 sq ft larger and has a garage your subject will not have, so it is superior on both. Subtract twice.
$276,000 − (50 × $60) − $12,000 = $261,000
Comp C is 50 sq ft smaller, and it has a half bath your subject will not have. Add for size, subtract for the bath.
$267,000 + (50 × $60) − $9,000 = $261,000
The market conditions over the nine months since Comp C closed showed no measurable movement in the county's own sold data, so it carries no time adjustment, and that absence is a decision you have to be able to support the same way you would support an adjustment.
Reconcile. Adjusted, the three sales say $258,000, $261,000 and $261,000. That is a $3,000 spread, which is tight. Comp A needed exactly one adjustment, which makes it the strongest single piece of evidence here, and it reads at the bottom. So the defensible ARV is $258,000 to $261,000, and the figure to underwrite is the bottom of it, $258,000.
Taking the bottom of the range is our own choice, not an appraisal convention. An appraiser reconciles to one supported opinion of value. Someone about to put money into a house has a reason to lean low.
Why the order matters, in dollars
Run the same three sales the other way. Average the raw price per square foot first, adjust after.
- Comp A: $255,000 ÷ 1,500 = $170.00 per sq ft
- Comp B: $276,000 ÷ 1,600 = $172.50 per sq ft
- Comp C: $267,000 ÷ 1,500 = $178.00 per sq ft
The average of those three is $173.50 per sq ft, and 1,550 × $173.50 = $268,925.
That is $10,925 above the adjusted, reconciled number, and the gap is not rounding. Price per square foot cannot see a garage or a half bath. Comp B's garage and Comp C's extra half bath got baked into the rate, and then the rate got multiplied across all 1,550 square feet of your house. Averaging raw price per square foot quietly charges you for features your property does not have.
Carry that error one step downstream. The 70% rule is a rule of thumb and no primary source sets it, but it is where most operators start, and ARV × 0.70 on the two figures gives $180,600 against $188,247.50. You would be about $7,600 high on the offer before you had even looked at the repair list.
The four mistakes that blow up a deal
- Averaging raw price per square foot and adjusting afterward. The example above is the whole argument. Our own coursework calls this the slip that blows a beginner's ARV, for exactly this reason. Adjust each comp to the subject first, every time.
- Treating a listing or a pending sale as one of your comps. A house listed at $300,000 tells you what one seller hopes for. Sold prices are facts, and you can usually pull them free from the county, see how to verify a property free. Fannie allows listings and contract offerings as supporting data and nothing more.
- Widening the market area until the comps agree with you. Reaching one school district or one nicer street over inflates the ARV and quietly eats the margin. Fannie's rule against expanding the neighborhood boundary exists precisely because it is tempting, and FHA's "not sales price" line is the same warning from the other direction. If the tight screen gives you three weak comps, widen the margin, not the map.
- Being optimistic. An optimistic ARV makes a bad deal look good, and it is the error that compounds, because every downstream number is a percentage of it. Be conservative on the ARV and conservative on the repairs, and let the deal be the thing that has to prove itself.
Why ARV feeds everything else
Once you have an ARV you can work backward to a maximum offer. The convention most operators start from is (ARV × 0.70) − repairs, and it is worth saying plainly that no primary source sets that 0.70. It is a rule of thumb, and the 30% it holds back is not margin. It is absorbing closing costs on both ends, the commission on the sale, holding costs while the work happens, financing costs, and the gap between your repair estimate and the repair bill.
Get the ARV wrong and every one of those is wrong with it. That is why learning to run comps and to estimate repairs are the two skills underneath everything else a wholesaler or a flipper does.
The come-up move
ARV isn't guesswork, it's a routine you can run the same way every time. At least three closed sales in one market area. Each comp adjusted to your property. Then reconcile, weighting the comps that needed the least correction, and underwrite the low end of the range. Do it in that order and your offers get sharper, and you can show anybody where the number came from.
Start free on Squatters to practice ARV and max-offer math on real off-market properties with our deal tools. Squat it. Fund it. Own it. 🦝