The deal analysis you do at the dining room table when the seller is sitting across from you bears very little resemblance to what your contractor finds three weeks later when they open up the walls. Most operators' underwriting is wrong, not because the math is wrong but because the inputs are wrong. The ARV is too high because the comps look comparable on paper and were not. The rehab budget is too low because the operator priced like they themselves were going to swing the hammer. The hold cost is too low because nobody planned for the months between "almost done" and "actually closed." The offer goes in, the deal closes, and three months later the operator is staring at a $32,000 budget gap they did not see coming.
This is a guide to the part of the business that decides whether the rest of it matters. Comp selection, adjustments, rehab scoping, the seven numbers that actually decide a flip, BRRRR exit math when rates move during the project, the hold-cost calculator nobody builds, and the operator discipline that separates an offer that closes profitably from one that becomes a slow-motion loss. Written from the operator side after 500+ acquisitions, an internal underwriting system we built because no off-the-shelf tool was honest enough about the things we cared about, and a lot of conversations with contractors that started with "yeah this is going to be a little more than we thought."
We will cover, in order: why most deal analysis is wrong, ARV comping on an off-market property (where MLS comps fail), the five appraisal-style adjustments that matter, the five-tier rehab system we use and what each tier really costs, the seven numbers that decide a flip, BRRRR exit math when rates move, the hold-cost calculator template, the contractor reality check, the city-vs-county adjustment that catches almost everyone in our market, and the seven mistakes we see operators repeat on underwriting.
Why most operators' deal analysis is wrong
The pattern: an operator looks at a property, pulls comps from a tool, runs the 70% rule in their head, calls the seller back with a number, and signs a contract. Three months later something has gone badly. We have seen the pattern enough times that the failure modes cluster into about four buckets.
Bucket 1: ARV anchored to bad comps. The comps were too old, too far, too different in size, too different in condition, or sourced from a non-renovated sale that drove the comp down. The ARV is off by 8-18%. On a typical Baltimore-metro flip that is $25-60k of margin gone before the project starts.
Bucket 2: Rehab budget priced like the operator was doing the work themselves. Operator who can swing a hammer estimates $32k for a kitchen, bath, paint, floors. Real contractor cost: $48k. The operator was pricing labor at their own opportunity cost (zero), not at the contractor's actual rate. Add 30-50% to the budget before you write the contract.
Bucket 3: Holding costs ignored or wildly underestimated. Operator budgets 4 months of hold; project actually takes 7 months. Insurance, utilities, taxes, debt service for 3 extra months on a $200k asset: another $8-15k of margin gone.
Bucket 4: Closing costs and resale costs left out entirely. Seller-side closing on the resale, agent commission, title fees, transfer taxes — operator did the math like they were going to FSBO. They were not. Another 6-9% of ARV.
Each bucket on its own is recoverable. All four together, on a single deal, turns a $40k projected spread into a $5k loss.
The fix is not better software. The fix is a discipline that forces the operator to be honest with themselves about each of the four buckets and to staple the math to the contract.
ARV — how to comp an off-market property
ARV is the most leveraged single number in the analysis. Get it right and a marginal deal works. Get it wrong and a great-looking deal is a trap. The mistake first-time operators make is treating ARV as a value the comping tool produces. ARV is a judgment call that uses comps as inputs.
Here is what an ARV comp actually has to look like in 2026 to be usable.
- Sold, not listed. Active listings tell you what someone thinks something is worth, not what someone paid. Sold-only.
- Within 90 days. 180 if you cannot find three within 90. Markets shift fast enough that older sales are increasingly noise.
- Within half a mile. Wider in rural markets. Tighter in dense urban markets where comps half a block over can be 40% different.
- Same property type. Single-family compares to single-family. 2-4 unit to 2-4 unit. Mixing is a category error.
- Renovated. This is the one most operators miss. The ARV is the value after your renovation. Comps need to be properties that were renovated to a similar level, sold by a flipper or by a meticulously-maintained owner. A 1960s house that has been lived in by the same owner since 1972 is not an ARV comp; it is an as-is comp.
- Same general condition profile. A fully gut-rehabbed property in a hot zip is not directly comparable to your medium-rehab project in the same zip.
- At least three. Two comps is not enough. Five is healthier. Use the median, not the average — averages drag toward outliers.
The tool you use to pull comps matters less than the discipline you apply to which comps you keep. We use DealCheck for the initial pull because the workflow is fast and the data is decent, but every operator we know who underwrites professionally also pulls the same property through BrightMLS or the equivalent for their market and reconciles the lists by hand. A pure tool-output comp set without operator review is the most common source of bad ARV.
The five appraisal-style adjustments that matter
Comps are rarely identical to your subject property. The five adjustments that matter, in rough order of magnitude:
1. Gross living area (GLA). Adjust per square foot, calibrated to the market. For Baltimore metro: $80-110/sf for the GLA difference, depending on the neighborhood. A subject property at 1,650 sf compared to a comp at 1,400 sf adjusts +$20-28k.
2. Basement. Finished below-grade square footage values at 50-70% of above-grade $/sf in our market. Finished with a full bath and egress: closer to 65-70%. Finished without egress: 50-55%. Unfinished basement: usually 5-15% of GLA $/sf, sometimes nothing at all. Never count unfinished space as living area; the comp adjustment is small but not zero.
3. Bedrooms / bathrooms. Extra bedroom: +$8-15k in Baltimore metro on a single-family. Extra full bath: +$5-10k. Extra half bath: +$2-4k. Adjust against the comp, not against your wish list.
4. Garage / carport. Detached garage: +$8-15k. Attached garage: +$10-20k. No garage where comps have one: subtract proportionally.
5. Condition / finish level. This is the one operators most often skip and it matters most. If your subject will be renovated to a "nice flip" finish and your comp was renovated to a "luxe flip" finish (engineered hardwood, quartz, designer fixtures), you cannot use that comp at face value. Adjust 5-12% down depending on the gap. Conversely, if your subject will be a clean medium rehab and the comp was a full gut with high-end finishes, the comp is not even in your category — drop it.
A good adjustment process produces a tight band of post-adjustment values. If your comps range $260k-310k and adjust to $278k-292k after the five above, your ARV is somewhere in the $285k area. If your comps adjust to $240k-320k after the five, your comp set is bad and you need to find better.
Rehab estimation — the five-tier system
Rehab is the second most-leveraged number, and it is almost always understated by inexperienced operators. We use a five-tier system internally because it forces honesty about scope.
Tier 1 — Light cosmetic. Paint, basic floor refresh, light fixture swaps, surface-clean kitchen and bath. Property is structurally sound, mechanicals are recent, no major systems failing. Baltimore metro cost: $12-22k for a 1,400-1,800 sf single-family.
Tier 2 — Cosmetic with mechanicals. Above plus one or two mechanical updates: replace the HVAC, replace the water heater, partial electrical update. Carpet replacement. Updated bath vanities and fixtures. Cost: $25-40k.
Tier 3 — Medium renovation. Kitchen replacement (new cabinets, counters, appliances), 1-2 bath replacements, full paint and floor, mechanical updates, some structural touch-ups. Cost: $50-75k.
Tier 4 — Heavy renovation. Above plus roof replacement or significant repair, full window replacement, electrical service upgrade, plumbing repairs, basement work, exterior siding or stucco. Cost: $85-130k.
Tier 5 — Gut rehab. Strip to studs. Full new mechanicals, full new layout possible, full kitchen and baths, new windows, new roof, new HVAC. Cost: $130-220k.
The numbers above are Baltimore metro rough averages, weighted toward Baltimore County. Baltimore City is generally +10-20% on the same scope because of the row-home structural issues (lead paint, plaster walls, knob-and-tube remnants, foundation movement). Other markets will be different — adjust by the labor and material indexes for your area.
The default for the operator: when in doubt about tier, assume one tier worse than the seller described it. The seller said "needs some paint and carpet." Walk in expecting Tier 2. The seller said "needs minor updating." Walk in expecting Tier 3. Sellers genuinely do not see what contractors see. They have lived with the issues for years.
A subset of operators try to refine the rehab budget after a quick walkthrough by themselves. This is a mistake at scale. The right discipline: walk through, identify the tier, apply your tier-based budget, add 15-20% contingency, and refine only after a real contractor has been on site with a scope of work in hand. Underwriting precision below 15% is false precision.
The seven numbers that decide a flip
Every flip's profitability is decided by these seven numbers. Get any one wrong and the deal compresses or fails.
- ARV. Most leveraged. Wrong by 5% destroys margin.
- Rehab budget. Second most leveraged. Wrong by 15% destroys margin.
- Purchase price. Determined by 1-2 backing into your minimum acceptable spread.
- Hold cost. Property taxes, insurance, utilities, debt service, monthly. Estimated months held: be realistic.
- Closing costs on acquisition. 2-4% of purchase, depending on the deal structure.
- Closing costs on resale. 7-9% of ARV: agent commission (5-6%), title (~1%), transfer taxes (varies by state — Maryland is ~1.5% combined for the buyer and seller, depending on the county).
- Financing cost. Points + interest on the hard money loan. Typically 3-5% of purchase + 12-14% APR on the portion actually drawn. See the hard money pillar.
The math is: Profit = ARV - Rehab - Purchase - Hold - Closing(acquisition) - Closing(resale) - Financing.
A worked Baltimore example. Subject: 3-bed 1-bath row home in a B-tier neighborhood. ARV after Tier 3 rehab: $295,000. Rehab budget: $62,000 with contingency. Hold cost: $1,900/month × 6 months = $11,400. Acquisition closing: $3,500. Resale closing: $24,000 (8% of ARV). Financing: $7,500 in points + ~$11,000 interest = $18,500.
Maximum allowable offer (MAO) at a $35k target profit:
- ARV: $295,000
- Minus rehab: $233,000
- Minus hold + closings + financing + profit: $295,000 - 62,000 - 11,400 - 3,500 - 24,000 - 18,500 - 35,000 = $140,600
So MAO is roughly $140,500. If we cannot get the property at that number, we walk.
This is the math. Every flip you analyze has the same seven numbers and the same equation. The discipline is to do the math, not to skip steps in your head because the deal "feels right."
BRRRR exit math when rates move
For BRRRR exits — buy, rehab, rent, refinance, repeat — the math adds two more numbers and one big risk.
The two added numbers are stabilized rent (what the property will rent for once renovated) and refinance LTV (what the DSCR lender will lend against the post-rehab value).
The risk is rate movement during the project. Hard money locks at one rate; the DSCR refi happens 6-12 months later at whatever rate exists then. If rates move 100-150 basis points during your project, your refi loan amount can drop because the property no longer DSCRs at the higher rate.
Worked example. Property: $200k purchase, $50k rehab. After-rehab value: $315k. Stabilized rent: $2,400/month. Hard money: 14% interest, 6 months. DSCR refi target: 75% LTV at projected 7.5% rate.
At project start, the math: $315k × 75% = $236,250 refi loan. Project total cost (purchase + rehab + interest) ≈ $263,000. Cash left in the deal post-refi: $263,000 - $236,250 = $26,750.
Now move rates +150 bps during your project. DSCR rate at refi: 9.0%. At 9% on a 30-year DSCR loan, the property's DSCR ratio at $2,400/month rent might not support 75% LTV anymore — the lender drops you to 70%. Refi loan now: $315k × 70% = $220,500. Cash left in: $263,000 - $220,500 = $42,500.
Same property, same rent, same renovation. $16,000 more cash trapped because rates moved during the project. The BRRRR margin compressed by half on a 150 bp move.
The defense is to model the refi math at rates 100 bps higher than current. If the deal still works at +100 bps and the cash-left-in is still acceptable, do the deal. If it does not, the deal is rate-fragile and you should reconsider or wait.
The other defense is to time the refi as early as possible. The shorter the gap between hard money origination and DSCR refi, the less rate exposure. Operators who hold open construction loans for 9 months "to make sure everything is settled" are taking a real rate risk for marginal benefit.
The hold-cost calculator nobody builds
Every operator we audit has a rehab number. Most have an ARV number. Almost none have a written hold-cost number with the inputs broken out. They estimate "5-6 months of hold" mentally and move on.
A real hold-cost calculator has these line items, per month:
- Property taxes (annual / 12)
- Insurance (vacant-dwelling rider is meaningfully higher than owner-occupied insurance — budget for the bump)
- Utilities (electric + water minimum, gas if applicable)
- Debt service interest on the drawn portion of hard money
- HOA/condo fees if applicable
- Lawn / snow / minor exterior maintenance during the project
- Theft or vandalism insurance deductibles, prorated as a small monthly charge for risk
For our Baltimore-metro hold-cost calculation on a typical single-family flip, this sums to $1,700-2,200/month. Project running 6 months: $11,000-13,000 of hold. Project running 9 months: $16,000-20,000.
The mistake operators make is budgeting for the expected duration of the project. Budget for the realistic duration. Inspectors are slow. Permits get held. Contractors take vacations. Materials arrive late. The number of operators who close a flip in the timeline they originally planned is small. The right discipline: estimate the timeline honestly, then add 30%.
For a project you think will take 4 months, budget for 5.5. For one you think will take 6, budget for 8. The math on the deal needs to work at the longer timeline, not the shorter one. If it only works at the shorter timeline, you have built fragility into the deal that will be tested.
The contractor reality check
The single largest gap between novice and experienced underwriters is the contractor reality check.
Novice underwriter: walks through the property, makes mental notes, estimates the rehab budget in their head. Reaches the offer in 20 minutes.
Experienced underwriter: walks through the property with a basic structural and mechanical checklist, identifies the tier, applies the tier budget, marks down five to ten specific items they could not assess and would need a contractor opinion on, adds 15-20% contingency, delivers an offer with the contingency baked in, and brings a contractor on site within the inspection period to refine the budget downward (or walk if the contractor opinion is meaningfully worse than the underwriting).
The discipline is to assume your underwriting is wrong by 15% in the worse direction until proven otherwise. The contingency is not optional. The contractor walk-through inside the inspection period is not optional.
What the experienced underwriter is looking for during the walkthrough is the items that turn a Tier 3 rehab into a Tier 4 rehab. These are the eight items that most often cause tier upgrades.
- Roof. If the roof is past 18 years of life or has visible curl/missing shingles, budget for replacement. $8-15k for a typical Baltimore row home.
- HVAC. Past 15 years of life, or visibly rusted. Replacement $5-9k for a typical system. Heat pump conversion: more.
- Electrical service. If the panel is a Federal Pacific or Zinsco, or under 100 amps, plan for upgrade. $2-4k for a service upgrade, $4-8k for a full rewire on an older home.
- Plumbing. Visible galvanized supply lines or cast-iron stacks at end of life. Stack replacement $4-8k. Full repipe $8-15k.
- Foundation. Visible cracks more than a hairline, particularly horizontal cracks or stair-step cracks at corners. Get a structural opinion. Cost can range from $3k cosmetic to $50k full underpinning.
- Sewer line. For older Baltimore row homes especially. Camera the line if there is any sign of slow drains or backyard movement. Replacement: $7-15k.
- Lead paint. Pre-1978 properties in Baltimore City and Baltimore County require lead-paint testing and remediation in rental scenarios. Budget for testing ($300-500) and for remediation if needed ($3-12k).
- Knob and tube wiring. In homes built before 1950, you may find remnants. Insurance and resale both penalize knob-and-tube. Budget for removal if found.
The walkthrough that checks these eight items in 15 minutes is the difference between an underwriter who delivers the project on budget and one who is constantly explaining cost overruns. None of this is hard to learn; it just has to be done every time.
The city-vs-county adjustment
This is specific to our market and similar in others. In Baltimore, properties in the city versus the county trade at meaningfully different ratios for the same nominal value. Same square footage, same bedroom count, same finish — a Baltimore City property and a Baltimore County property of identical specifications can have ARVs that differ by 12-20%.
The cause is a combination of: city school district reputation, city property tax rates (significantly higher than county), perceived crime data, code-enforcement aggressiveness, lead-paint requirements, the higher general cost of ownership on a city row home vs a county detached single-family.
The operator implication: when underwriting a Baltimore City property, take 10% off whatever ARV the tool produces, then re-comp manually using city-only sales. The tool is averaging across geographies; you cannot.
The same dynamic exists in many metros — center-city vs near-city-suburb vs further-suburb. Always identify the specific geographic micro-market your subject is in, and find comps only within that micro-market.
The seven mistakes we see operators repeat on underwriting
In rough order of frequency.
1. Using as-is comps instead of renovated comps. Most common single error. Operator looks at "what sold nearby," does not distinguish between renovated flips and tired owner-occupied sales, gets a comp set that drags ARV down (or up) for the wrong reasons. Fix: filter comps to renovated only. If the comp was not renovated or recently rehabbed, it is not an ARV comp.
2. Self-pricing rehab. Operator who is handy estimates the rehab at the price they would do it themselves. They are not doing it themselves. Fix: price labor at the contractor rate. If you do not know your local contractor rate, learn it before you write your next offer.
3. Ignoring contingency. Underwriting without a 15-20% rehab contingency is gambling. Every project has unknowns. The contingency is the budget line for those unknowns. Fix: bake it in. Make it visible. Treat it as inviolable.
4. Forgetting resale closing costs. Operator underwrites against ARV gross, not ARV net of resale closing costs. 7-9% of ARV disappears when you sell. Fix: model net ARV, not gross.
5. Project timeline optimism. "It'll take 4 months." It takes 6. The hold cost on the extra 2 months eats $4-6k. Fix: add 30% to your honest estimate.
6. Single-comp anchoring. The operator finds one beautiful comp at $310k and anchors to it. The other three comps adjust to $278k. The MAO computed off $310k is too high by $16k. Fix: median of adjusted comps, never single-comp anchor.
7. Skipping the contractor walkthrough. The operator writes the offer, the seller signs, and the operator never brings a contractor on site during the inspection period. By the time the contractor sees the property, the operator is past the contingency and on the hook for whatever the contractor finds. Fix: contractor walkthrough within the inspection period, every time. Make it a stage gate.
The minimum viable underwriting system
If you are looking at this guide because you do not yet have an underwriting discipline, the minimum:
- A comp-pulling tool you trust (DealCheck, PropStream, or your MLS access) — pick one and use the same one for every deal so your discipline is consistent
- A written rehab tier system calibrated to your specific market — the five-tier model above is a starting point; refine with your contractor relationships
- A hold-cost worksheet that itemizes the seven inputs above
- A walkthrough checklist that covers the eight items in the contractor reality check section
- An offer template that bakes in 15-20% rehab contingency
- A discipline that the contractor walkthrough happens inside every inspection period, no exceptions
That is the system. It will not produce a perfect ARV — no system will — but it will produce a defensible one. The operators who underwrite with a system like this close at 8-12% margin reliably. The operators who underwrite by feel close some deals profitably and lose some, and over time it averages out to break-even.
For neutral, operator-written reviews of the underwriting and comp tools we have evaluated, see the analyze-deals category. For the funding side that determines your financing constraints, the hard money pillar. For the lead side that determines whether you ever even see a deal worth underwriting, the lead-stack pillar. For what happens after the underwriting and the contract — moving the deal to a buyer — see the CRM + dispo pillar.
The math on a flip is unforgiving but learnable. Most of the operators we see fail at underwriting fail not because the math is hard but because they did not do the math. They estimated, they trusted the tool, they trusted the seller's condition description, they trusted their own gut. The math is cheap. It costs you 20 minutes per deal. The discipline of doing it every time, on every deal, against the seven numbers and the five adjustments and the five tiers, is what separates an operator who is in business in five years from one who is not.