Purpose
Turn a subject property's attributes into a ranked, defensible set of comparable sales — scored on location, physical similarity, and transaction recency — with per-comp adjustment guidance and a written rationale that holds up in an appraisal review or investment committee. Eliminates cherry-picking and the "gut-feel comp" problem that inflates or deflates valuations.
When to Use
Invoke this skill when the user asks to:
- Find or rank comparable sales for a CMA or broker price opinion
- Select comps for acquisition underwriting or an offer decision
- Build a defensible comp set to support or contest an appraisal
- Screen a raw pool of recent sales down to the 3–6 best comps
- Justify why specific properties were chosen or rejected as comparables
- Estimate a value range for a property using recent transactions
Inputs Required
Before starting, collect:
- Subject property attributes: Address/location, property type, GLA (square footage), bed/bath count, lot size, year built, condition, and any premium features (view, garage, pool, updates)
- Comp pool: The candidate sales to evaluate — address, sale price, sale date, and matching attributes for each (paste a list, table, or MLS export)
- Valuation purpose: CMA / appraisal support / acquisition underwriting / dispute — determines rigor and adjustment conservatism
- Effective date of value: The date the valuation is anchored to (governs recency scoring)
- Market condition: Appreciating / stable / declining, plus rough monthly rate if known
- Geographic constraint: Acceptable radius and any hard submarket boundaries (school zone, HOA, waterfront line)
If market condition or appreciation rate is unknown, assume a stable market and flag any time adjustment as an assumption to verify. If the comp pool is not provided, ask the user to supply candidate sales before proceeding — this skill ranks a provided pool, it does not fabricate transactions.
Instructions
Step 1: Establish the Comparability Standard
Define the subject's non-negotiable match criteria before scoring anything:
- Property type: Same use class (SFR to SFR, not SFR to townhome unless the market treats them as substitutes)
- Location tier: Same submarket, school district, and HOA boundary where those drive price
- Transaction validity: Exclude non-arm's-length sales (family transfers, REO/foreclosure, short sales) unless the entire market is distressed — flag any distressed comp used
- Recency window: Prefer sales within 90 days of the effective date; extend to 6 months only if inventory is thin; note any comp beyond 6 months
Screen the pool against these gates first. Anything failing a hard gate is disqualified — record why.
Step 2: Score Each Surviving Candidate (0–100 Similarity)
Apply weighted similarity scoring across three dimensions. Default weights (adjust for property type and state them explicitly):
Location — 40 points
- Same immediate neighborhood/subdivision: full points; adjacent comparable area: partial; superior/inferior area: penalize with directional note
- Proximity to the same amenities/disamenities (view, arterial road, waterline, school)
Physical — 40 points
- GLA within ±10% of subject: full; scale down as the gap widens (comps outside ±25% GLA are weak)
- Bed/bath match, lot size proximity, year built within ±10 years, comparable condition and updates
Transactional — 20 points
- Sale within 90 days: full; scale down by recency
- Clean arm's-length transaction with normal financing/concessions
Produce a numeric score per candidate and a one-line reason for the score.
Step 3: Apply Directional Adjustment Guidance
For the top-scoring comps, indicate how the subject compares on each material difference and the direction of adjustment (do not fabricate precise dollar amounts unless local adjustment rates are supplied):
- GLA: Subject larger → adjust comp up; smaller → adjust down (state suggested $/sqft basis if provided)
- Condition/updates: Subject superior → comp adjusted up; note major systems (roof, HVAC, kitchen, bath)
- Lot / view / premium features: Directional adjustment with magnitude qualifier (minor / moderate / major)
- Time/market: Apply appreciation-based time adjustment to the comp's sale price if the market is moving Flag any comp requiring net gross adjustments large enough to weaken reliability (a common threshold: net adjustments over 15% or gross over 25% signal a weak comp).
Step 4: Rank and Select the Final Comp Set
- Rank all candidates by similarity score
- Select the strongest 3–6 as primary comps; the market's homogeneity dictates how many
- Designate 1–2 secondary/support comps that bracket the primaries
- For each primary comp, write a 2–3 sentence selection rationale tying it to the subject
- For the top 3 rejected candidates, write a one-line rejection reason (fails recency, superior location, GLA gap, distressed sale)
Step 5: Reconcile a Value Indication
- State the adjusted-price range implied by the primary comps
- Weight toward the highest-scoring, least-adjusted comps and explain the weighting
- Give a single point estimate or tight range with a confidence note (high / moderate / low) based on comp quality and dispersion
Output Format
- Structure: Comparability standard → scored candidate table → adjustment guidance → ranked final set with rationale → rejected comps → reconciled value
- Length: 700–1,100 words
- Format: A scored comparison table for all candidates, plus structured bullets for rationale and reconciliation
- Sections: Comparability Standard → Candidate Scoring Table → Adjustment Guidance → Final Comp Set (ranked, with rationale) → Rejected Candidates → Value Reconciliation
- Tone: Valuation analyst — evidence-first, cautious about precision, transparent about assumptions
Quality Checklist
Before delivering output, verify:
- [ ] Every candidate in the pool is accounted for — scored, or disqualified with a stated reason
- [ ] Similarity scores use stated weights and sum correctly to a 0–100 scale
- [ ] The final set contains 3–6 primary comps, each with a 2–3 sentence rationale
- [ ] Every distressed, non-arm's-length, or >6-month comp used is explicitly flagged
- [ ] Adjustment guidance states direction for each material difference — no unlabeled adjustments
- [ ] The value reconciliation includes a confidence level tied to comp quality and dispersion
Notes
- The single most common error is over-adjusting a poor comp instead of rejecting it — a comp needing large net adjustments is telling you it doesn't belong.
- Recency and location beat physical similarity in fast-moving markets; a slightly different house sold last month often beats a near-identical house sold a year ago.
- When the comp pool is thin, widen the recency window before widening geography — a distant comp in a different submarket is usually less reliable than a slightly stale local one.
- This skill ranks and reasons over a supplied pool; it does not invent transactions. If precise dollar adjustments are needed, request local paired-sales adjustment rates.
How to use this skill
This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.
Installing in Claude Code:
~/.claude/skills/<folder-name>/SKILL.md— auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)<your-project>/.claude/skills/<folder-name>/SKILL.md— auto-loads only when working in that project (recommended for project-specific skills)
Restart Claude Code or open a new session to register.
Using in Claude.ai, ChatGPT, Gemini, or other AI tools:
- Drag this file into a new chat, or copy and paste its contents
- Send your task or request as your next message — the AI will apply the skill where it fits
For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills