Methodology
Carfocal is built around transparent, citeable vehicle facts. We aggregate primary datasets and then publish derived metrics that help shoppers compare vehicles quickly.
Primary Sources
- NHTSA crash test ratings, recalls, and owner complaints.
- EPA fuel economy ratings (US-market).
Each guide page includes a Sources section and updated timestamp so readers can verify the underlying filings.
CarFocal Reliability Score
The v3 Carfocal Reliability Score is a 1.0–10.0 research metric based on severity-weighted recall campaigns, complaints adjusted for estimated sales exposure and vehicle age, service bulletins, and supported major-issue assessments. Higher values mean a lower calculated risk signal. It is not a measured probability of failure.
Current v3 calculation
- Exposure = sales-volume estimate × vehicle age, with age bounded to at least one year.
- Smoothed complaints per 10,000 exposure units = (complaints + 65.8443) ÷ (exposure + 340203.5972) × 10,000.
- Recall severity weights: critical/high 3; moderate 1; minor/low 0.25. When detailed severity is unavailable, the count is used.
- Risk = 0.5 × smoothed complaint rate + ln(1 + weighted recalls ÷ age) + min(0.6, 0.18 × ln(1 + TSBs ÷ age)) + major-issue adjustment.
- The major-issue adjustment is 0.6 × assessment confidence when a high-risk engine/transmission issue is identified.
- Score = 1 + 9 × exp(−0.15 × risk), bounded to 1–10 and rounded to one decimal.
Important: Complaint counts reflect reporting volume, not a guaranteed failure rate. They should be read as a starting point to investigate common issues and verify recall completion.
Sales exposure can use a model estimate, a brand-based proxy, or a fallback of 25,000; it is not necessarily the vehicle's measured production count. The calibration uses 6,430 model-year records. Missing counts do not establish that there were zero issues; the calculation currently uses zero when a count is unavailable. Safety ratings are a separate overall-score component, not an input to this v3 reliability formula.
Recall figures count campaigns for a model year and can span configurations. They do not show whether your individual car has an unrepaired recall. Check your VIN with NHTSA.
Updates and Freshness
Recalls and complaints can change over time. A report's update date records its content update; source retrieval dates are shown separately where available. An update date is not proof that every source was checked that day. Prices and failure-mileage ranges are estimates, vary by configuration and region, and should be checked against current service records and local repair quotes.
Global Rankings
Carfocal also publishes build-time research pages that aggregate manufacturer performance across eligible vehicles, both for all years combined and as yearly snapshots.
- Overall Score Index: average v2 buyer-guide overall score by manufacturer. Components are reliability (30%), safety (25%), maintenance (15%), efficiency (15%) and practicality (15%). Tuning is excluded. Available components are weighted, with missing-data blending toward 6.5, confidence adjustments and risk caps. EV maintenance can use a 7.5 prior; selected high-performance brands exclude efficiency and normalize remaining weights.
- Reliability Index: average vehicle reliability scores by manufacturer.
- Ownership Risk Index: a 50/50 blend of inverted reliability score and maintenance burden.
Each index also publishes yearly static reports at URLs like /manufacturers/rankings/reliability/2020. Index pages publish an in-page changelog and use the same scoring functions as vehicle reports. Brand scores average eligible model-year records, not sales-weighted fleet outcomes; differing coverage and age can affect comparisons.