Philosophy
We blend verifiable data with practical context. Signals are grouped so that similar ideas work together. For new vs. used, we emphasize different factors to reflect how buyers actually evaluate cars.
What we share: the intent of each group and the kind of data it considers.
What we keep private: exact formulas, thresholds, and internal weights.
What we keep private: exact formulas, thresholds, and internal weights.
How the score is built
Baseline
Every car starts from a neutral foundation.
Group Signals
Each group nudges the score up or down based on evidence.
Safeguards
We handle missing/uncertain data conservatively (no scary guesses).
Medal & Verdict
Tier medal + Lady’s plain-English summary.
Tie-Breaker
Price-per-Point highlights value when scores are close.
Scoreing at a Glance
▶🚗 Mileage vs. Age
Assesses usage patterns relative to age to contextualize wear. Outlier usage is treated as a risk or advantage signal.
▶🛡️ Safety Signals
Accounts for meaningful safety content appropriate to the model year and class, prioritizing verifiable features.
▶💵 Price Context
Benchmarks asking price against comparable vehicles to reflect deal quality; extreme divergences trigger caution.
▶🔧 Condition Signals
Aggregates available indicators of care and wear to inform expected ownership experience; avoids speculation when data is sparse.
▶🌨️ Drivetrain & Region Fit
Evaluates traction configuration against typical regional conditions to capture real-world usability.
▶☀️ Climate Features
Recognizes feature sets that materially improve comfort and safety in local weather scenarios when known.
▶🐾 Pet/Kid-Friendly Vibes
Highlights practicality factors valued by pet owners (load area, ventilation, cleanability) that influence day-to-day utility.
▶💸 Ownership Cost Signals
Synthesizes broad indicators tied to ongoing cost expectations to reduce total-cost-of-ownership surprises.
▶⚠️ Recall Awareness
Reflects recall awareness when supported by authoritative sources; remains silent where data is unavailable.
▶🔄 Turn Over
Interprets owner count in context; pattern and interval quality matter more than raw totals.
▶⚙️ Powertrain Suitability
Checks that performance characteristics align with intended role and class instead of rewarding speed alone.
▶🔌 EV-Specific Considerations
Applies EV-only context (e.g., practical range framing, charging suitability) where data support exists.
ℹ️
Investor Notes:
- Evidence-Weighted: Inputs are verifiable where possible; missing data is handled conservatively (no invented values).
- Modular & Extensible: Groups are independent “signal families,” enabling targeted updates without platform churn.
- New vs. Used Awareness: Internal routing emphasizes different signals by vehicle lifecycle to protect consumer clarity.
- Consumer Transparency Layer: Lady’s summary explains the why without exposing IP; Price-per-Point resolves near ties.
- Compliance & Privacy: We avoid storing or quoting user-supplied third-party reports; interpretation is session-scoped.
What Lady Explains
Lady’s verdict always leads with Year / Make / Model / Trim (and drivetrain when known), then summarizes the “why” behind the score in plain English. Her tone reflects the result: more excited for higher scores, more cautious for lower scores.
- No copy/paste from third-party reports; if you upload a report, Lady interprets it in the moment without storing or quoting the source.
- When data is missing, Lady avoids making claims and sticks to what we know.
Fairness & Transparency
- No secret penalties for missing info: missing ≠ bad. We don’t punish a car for data a seller never provided.
- Price-Per-Point: when two cars are close, we show which one delivers more score per dollar.
- Context beats absolutism: a feature that’s essential in Minnesota might be optional in Miami—signals respect locale.
Questions?
We love curious users. Ask anything about our approach (we’ll answer without exposing internal math). If something looks off, flag it—feedback helps us improve.