How Alexander Atlas Scores Markets

Scoring Formula — AA 4.0

Atlas Score™ is a 100-point composite index. Each market is scored across 12 dimensions (the 12th, Desirability, is used only by the Owner-Occupant lens); raw metric values are transformed into percentile ranks against the full NC + SC distribution, then mapped to points within each dimension's weight cap. V3 adds two new categories — Climate & Insurability (sourced from FEMA NRI county hazard data) and Market Opportunity (HUD Opportunity Zone + STR demand proximity + HMDA cash buyer %) — plus a macro regime multiplier tied to the 10-year Treasury rate (FRED DGS10) and compound interaction bonuses for aligned forward signals. Safety & Quality carries no weight on the balanced and investor lenses — it is shown there for transparency — and is a primary scored category on the Owner-Occupant lens. Crime is scored only in that category; it is not folded into Risk Profile or any other dimension.

How to read six scores

• The Composite is its own score, not an average of the other five. It is a balanced view in its own right. A market can beat its Composite on one lens and fall well below it on another.
• ★ Best Fit is a description, not a recommendation. It marks a market's highest-scoring lens — which strategy the market suits best, not that the market is good. A market can be Best Fit for Short-Term Rental and still score in the 40s. Read the number, not just the badge.

V3.2 — Anchor Access. Scoring now incorporates distance-decayed access to the Carolinas' major demand anchors: metropolitan cores, commercial airports, and major universities. These are access signals — they credit reach to deep buyer and tenant pools, bounded within existing dimension caps. Where an anchor's demand is already priced into a market's own fundamentals (for example, campus-core renter rates), the access signal is suppressed to prevent double-counting — and remote markets are never penalized for distance. The anchor tables, decay curves, and caps are part of Alexander Atlas's proprietary engine; the nearest anchors and distances for any ZIP are cited directly in the map panel and property reports.

V3.3 — Live Market-Data Signals. Five additional signals refine the score without changing the 100-point structure or the dimension weights. Each is a bounded, presence-guarded adjustment applied inside an existing dimension, so a market with none of the new data scores exactly as before:

• Closed-sale price trend (ATTOM) grounds the appreciation read in actual recorded sale prices rather than listing estimates. Where recorded sales diverge from the listing trend, the recorded figure is given weight.
• New-business formation (BLS QCEW county establishment & employment growth) feeds Economic Strength as a leading indicator of jobs and rental demand.
• In-mover income quality (IRS SOI — average adjusted gross income of in-migrating households) informs Demographic Momentum: it distinguishes high-income in-migration from low-income, not just the count of movers.
• Insurance premium trajectory & carrier withdrawal (state rate-filing exposure tiers) feeds Climate & Insurability as a carrying-cost and financing headwind in exposed coastal markets.
• Distress / long-term vacancy (Census ACS other-vacant rate) informs Risk Profile and buyer leverage as a motivated-seller signal.

A separate data-confidence discount pulls markets with genuinely thin source coverage toward a conservative anchor, so a data-sparse market cannot read as falsely confident; well-covered markets are untouched. The exact adjustment magnitudes, thresholds, blend ratios, and confidence bar are part of the proprietary engine.

AA 4.0 — Buyer-Type Reprice (July 2026). The five strategy profiles were re-derived against how each strategy is actually underwritten, and every scored market in North and South Carolina was repriced on the same day. The 100-point structure and the dimension list are unchanged, and the balanced Atlas Score™ moved least. What changed is how the strategy lenses read those same dimensions:

• Rental / Hold was rebuilt around debt-service coverage. The previous profile tracked a market's home values more closely than the yield metrics a hold is actually financed on. The current one tracks debt-service coverage, cap rate and gross yield far more closely — and no longer rates a market investment-grade when the rent does not cover the debt.
• Owner-Occupant became a full lens rather than an investor profile with the yield turned down. Schools, safety and neighborhood stability are scored, and yield is not.
• BRRRR carries a heavier risk weight. It is the one strategy defined by leverage, and leverage amplifies downside by construction.
• Flip and STR were held at their prior values, so movement on those lenses reflects the market, not the model.

Versioned score history. Every recorded snapshot now carries the methodology version that produced it. When a score moves because the model changed rather than because the market did, Atlas Score History™ draws that step in slate and labels it a model update — distinct from the gold marker used for a genuine grade change. History is never back-filled: a date that was not scored never acquires a point.

Data Sources

Every metric is sourced from a named, public or licensed dataset with a documented vintage date. We do not use proprietary black-box estimates.

Percentile Ranks

Raw scores are also expressed as percentile ranks — computed by comparing each market's score against the full distribution of all 1,101 scored NC and SC markets.

Three ranks are shown per market: overall (all 1,101 markets), state-level (NC or SC only), and county-level. This lets investors compare a market to both its local peers and the full two-state universe.

Buyer-Type Investment Modes — shown as “Strategy Fit” in the app

Beyond the balanced Atlas Score™, every market is scored five more times — once per strategy — by re-weighting the same components. In the live market map these are the Strategy Fit cards (Flip · Rental/Hold · BRRRR · STR · Owner-Occupant), each with its own 0–100 score and A–F grade. Every profile still sums to 100, so the five scores stay directly comparable to the balanced score and to each other. The Owner-Occupant lens is the one built for people buying a home to live in rather than to let: it scores safety and desirability heavily and ignores yield entirely.

The matrix below shows the driver emphasis each strategy places on every component. The precise point weights are part of Alexander Atlas's proprietary scoring engine; what matters for reading a Strategy Fit score is which components drive it.

How to read the profiles: Flip leans on appreciation, demographic momentum and supply constraint — it’s resale-driven. Rental / Hold leans hardest on cash flow and rental demand — pure yield. BRRRR leads on cash flow and rental demand with appreciation supporting the refinance — refinance-and-hold. STR is dominated by Market Opportunity (short-term-rental demand proximity) plus cash flow. Safety & Quality is display-only on the balanced and investor lenses, and a primary scored category on Owner-Occupant. It is never folded into Risk Profile.

Each buyer-type score uses the same 0–100 scale and A–F grade bands as the balanced Atlas Score™. The “★ Best Fit” badge on the market map marks the strategy with the highest score for that market.

Market Timing & Buyer Leverage V3.1

Alongside the quality-focused Atlas Score™, every market now carries a forward-looking market-timing layer. It answers a different question — not "how good is this market?" but "is now the time to buy, and how much room do I have to negotiate?" These signals appear as the buyer/seller badge on each ZIP, power the "Buyer / Seller Markets" filter in the dashboard, and surface in Atlas Report™ property reports as a page-1 dual verdict plus a Negotiation Leverage block on the offer page.

AA 4.0 — the growth gate now floors rather than zeroes. A market with soft population growth could previously have its timing credit switched off entirely, which suppressed the signal in precisely the high-buyer-leverage markets it exists to surface. It now scales down to a floor instead, so buyer leverage is always reflected in the timing read.

Why this matters: a cold market can mean a temporary buyer's window in a growing area (a deal) or a structurally declining one (a trap). The Market Balance Index alone cannot tell them apart — Entry Leverage crosses market softness with demand growth so the two are never confused. In the dashboard's Buyer / Seller Markets view, ZIPs are ranked by Entry Leverage so genuine early-entry deals surface above value-trap deep-discount markets.

Grade Bands

Property-Level Atlas Score™ (Atlas Reports)

When you order a full Atlas Report on a specific property, a separate 100-point property-level Atlas Score™ is computed. This score evaluates the individual deal — not just the market — across 7 dimensions. Like the market score, it is strategy-weighted: the seven dimensions are re-weighted for Flip, Rental/Hold, or Owner-occupant buyers, and the report leads with the best-fit strategy’s weighting (each column below sums to 100).

Property-Level Data Sources

An Atlas Report™ draws on address-level data the market score does not use. None of the sources below feed the Atlas Score™ — see the table above for that. Every rent signal is listed on the report, and material disagreement between the property-level models is flagged rather than averaged away.

How they are combined. An Atlas Report draws on address-level data the market score does not use. Valuation: closed comparable sales from ATTOM and Zillow, scored through a six-phase similarity and adjustment methodology that produces an A–F comp confidence tier; where qualifying comps are unavailable, ARV falls back to zip-level medians. Rent: the Zillow Rent Zestimate — or buyer-verified comparable leases where you supply them — is the underwriting basis, with the ATTOM Rental AVM, the RentCast long-term rent AVM and HUD Fair Market Rents collected alongside it. Every rent signal is listed on the report, and material disagreement between the property-level models is flagged rather than averaged away. Market-level (ZIP) scoring uses none of these — see the source table above for what feeds the Atlas Score™.

Scores ≥ 70 appear on reports as PROCEED, 50–69 as PROCEED WITH CAUTION, and < 50 as HIGH RISK. These are analytical outputs, not buy/sell recommendations — always perform independent due diligence before any transaction.

Known Limitations

• Crime data lags 1–3 years. FBI UCR data is published with a 12–24 month delay. When the FBI Crime Data Explorer API is unavailable, scores fall back to static NC DPS / SC SLED county anchors from 2023. Rapidly changing markets may have outdated crime scores.
• BLS unemployment and HUD FMR are county-level, not ZIP-level. All ZIPs within a county share the same unemployment rate and rent benchmark — intra-county variation is not captured.
• Census ACS margins of error. ACS 5-year estimates for rural ZCTAs (population < 2,500) carry wide confidence intervals. Treat income and population figures as estimates, not census counts.
• Thin Zillow markets. ZIPs with fewer than 5 annual sales may have missing or interpolated ZHVI/ZORI values. These markets are flagged as "limited comps."
• Scores reflect market trends, not individual property values. A high-scoring ZIP does not guarantee any specific property will appreciate or cash-flow positively. Property-level underwriting is required before any investment decision. Scores are not appraisals or valuations.
• Scores describe markets. Every input measures housing-market economics or public amenities. No score is a recommendation to buy in or avoid any area.
• Cap rate is modeled, not observed. The cap rate estimate applies county-level vacancy and operating expense assumptions. Actual cap rates depend on the specific property and financing structure.

Update Schedule

Questions or Corrections

If you spot a data error or have a methodology question, email support@alexanderatlas.ai. We take data accuracy seriously and will investigate and correct verified errors within 10 business days.