apartment-scorecard

Apartment Scorecard — Compare Rentals Like an Analyst

Overview

Apartment hunting corrupts judgment: you tour five places in a day, the one with the nice lobby light wins, and six months later you're paying $120/month in "amenity fees" for a gym you never use with a 50-minute commute you didn't count. The listing price is marketing; your actual cost and quality are computable.

This skill enforces the process your rational self would use: hard constraints first (no point scoring a place you can't afford, can't bring the cat to, or can't move into on time), then weighted scoring on 16 criteria with weights that reflect YOUR priorities (a remote worker weights light and noise; a commuter weights transit), then true monthly cost — every fee, amortized broker fee, parking, pet rent, utilities, insurance, deposit interest, and optionally your commute time valued in money — and finally negotiation prep: the market signals that justify asking for less, with anchoring numbers and a script.

When to Use

• Comparing 2+ listings you've toured (or are about to) → screen, then compare the finalists
• "Can I afford this rent?" → budget (the 30/33/50 rules + landlord 3×-income math)
• Touring today: score each place on the 1-5 criteria sheet right after leaving (memory decays in hours)
• About to sign: negotiate with the facts you gathered (days vacant, comparables, lease flexibility)
• Defining the search itself: hard constraints (max commute, pets, budget) before you waste tours
• Don't use for: buying property (mortgage/tax math differs — see job-offer-comparator style skills for offers), roommate conflict mediation, or short-term vacation rentals.

The 16 Criteria

Edit weights in ~/.apartment-scorecard.json — set to 0 anything you don't care about. Score each listing 1-5 per criterion during/immediately after the tour; criteria prints this table.

Commands

# Affordability: 30/33/50% rules, 3×-income approval math, front costs
python3 scripts/apartment_scorecard.py budget

# See criteria + defaults
python3 scripts/apartment_scorecard.py criteria

# Screen & rank: hard constraints kill, survivors score, true cost ranks
python3 scripts/apartment_scorecard.py screen --file apartments.json

# Side-by-side of finalists + quality-vs-cost premium
python3 scripts/apartment_scorecard.py compare "Maple St 2BR" "Oak Rd Garden"

# Negotiation plan with your leverage facts
python3 scripts/apartment_scorecard.py negotiate "Maple St 2BR" \
    --vacant-days 30 --lease-offer 18 \
    --facts comparables move_in_speed

# Full demo: budget → screen → compare → negotiate
python3 scripts/apartment_scorecard.py example

Listings: JSON list (or CSV) — name, rent, bedrooms, commute_min, deposit, fees_monthly, parking_monthly, pet_rent_monthly, broker_fee, utilities_included, pets_ok, available (YYYY-MM-DD), lease_months, plus your 1-5 scores per criterion. See references/listing-format.md.

True Cost Model

true_monthly = rent + monthly fees + utilities (if not included)
             + renters insurance + parking + pet rent
             + deposit × 4% / 12        (interest your money loses)
             + (broker + move-in fees) / lease months
             + commute_min × 2 × days × 4.33 × $/min (optional, try 0.3-0.8)

The commute term is the sleeper: at $0.50/min, a place 20 minutes further costs ~$433/month — often more than the rent difference people agonize over. Report also shows %income so the 50% danger line is visible at a glance.

Workflow

1. budget → know your real ceiling BEFORE touring (all-in, not rent).
2. Set hard constraints in the weights file; don't tour what fails them.
3. Tour with the criteria sheet; score 1-5 within an hour of each visit.
4. screen → ranking with hard fails listed; compare the top 2 → the premium line asks the right question ("is $300/mo worth 9 score points?").
5. negotiate the chosen one with real facts → anchor first, trade lease length/speed for dollars, get it in writing.

Common Pitfalls

1. Comparing rent instead of true cost. A $1,690 place with no fees and included utilities can be cheaper than a $1,580 place with $150/month of parking + amenity + pet fees. The screen table's true $/mo column exists to prevent this.
2. Scoring during the tour by vibes, then writing numbers later. Score within the hour; bring the criteria list; photograph everything (the photos double as deposit-dispute evidence later).
3. Ignoring the commute because "it's fine." Value your time honestly (the $/min setting); 30 extra minutes daily is ~22 hours/month.
4. Hard constraints as soft preferences. If pets are non-negotiable, they're a constraint, not a weight — mixed-up categories is how people end up heartbroken at lease signing.
5. Negotiating by asking "is the rent flexible?" That invites a no. Anchor with a specific number and a trade (longer lease, fast move-in, auto-pay); negotiate prints the script.
6. Forgetting front costs. Deposit + first + last + broker = 2.5-3.5× monthly rent due on signing day; budget it before you fall in love.
7. Weighting everything 3. If everything matters, nothing does. Force yourself to 0-out at least five criteria.

Verification Checklist

• python3 scripts/test_apartment_scorecard.py → ALL TESTS PASSED (29 assertions)
• python3 scripts/apartment_scorecard.py example runs the full pipeline
• Weights file edited to YOUR priorities (not defaults)
• Every toured listing scored within an hour of the visit
• Decision made on true $/mo + score, not advertised rent + vibes