How we score reliability

Last updated: October 31, 2026

Every score and label on motorwick comes from the same rule, applied the same way to every model. Nothing is hand-picked, paid for or adjusted for a brand. This page explains the rule and where it can mislead.

1. Count the complaints

We start from the NHTSA consumer complaint file, which holds every vehicle safety complaint owners have filed with the U.S. National Highway Traffic Safety Administration since 1995. For the 119 models we score, that is 660,206 complaints, with data through October 28, 2026. We count each complaint once per model year, even when it names several parts, and group it by system: engine, transmission, electrical, brakes, steering and so on.

We never publish what owners wrote, where they live or their VIN. Only counts appear on the site.

2. Divide by cars sold

A model that sells 400,000 cars a year will collect more complaints than one that sells 40,000, even if each car is just as good. So we divide each model year's complaints by the number of that model sold new in the U.S. in the same calendar year, and express it per 100,000 sold. Most cars of a model year are sold in that calendar year; the rest are sold in the months before, so the match is close but not exact.

The model average is the total of a model's complaints divided by the total of its sales, across all the years we score.

3. Label a year only when the numbers are clear

Most complaints: the year's score is more than twice the model average, and the year has at least 50 engine or transmission complaints or an engine or transmission recall or investigation. Both conditions must hold, so a year cannot be flagged only because of a few loud reports about a radio or a seat.
Fewest complaints: the three lowest-scoring years that are below the model average and at least three model years old. Newer cars have had less time to collect complaints, so they cannot be called the best yet.
Every other year gets no label. The score and the counts are still shown.

The words we use are "most complaints" and "fewest complaints", never "defective" or "safe", because a complaint is a report, not a finding.

When we do not score

No separate sales figure. Ford F-250 and Ram 2500: automakers do not publish these separately, so we show raw counts only and label nothing.
Sales published as a combined total. Chevrolet Silverado, Ford F-150, GMC Sierra, Ram 1500 and Toyota Prius: the sales figure covers related versions (for example a whole pickup line). Scores are still useful for comparing the model's own years, but these models stay out of rankings and class averages.
Very small years. A launch or run-out year with sales under 15% of the model's typical year is marked as a partial year and not scored.

Rankings and class averages

Rankings compare the average score of each model over model years 2015 to 2023, using only models with at least four scored years in that window and a clean sales figure. Class averages (for example compact SUVs) are built the same way, year by year.

What the numbers cannot tell you

Complaints are not verified by NHTSA. Some describe real defects, some describe normal wear or misunderstandings.
Owners of some brands and of newer, more connected cars may be more likely to file reports.
A big recall can trigger many complaints for a single year.
The data says nothing about one particular vehicle. Its maintenance history and an inspection matter more.

Updates

We import the NHTSA files and the EPA fuel economy file every month and the EIA gas price every week. If a new file is incomplete or its format changes, the import stops and the site keeps the previous data. Sales figures are updated each quarter. Every page shows the date of its data.

Questions or a correction: write to [email protected]. The full list of data sources and licenses is on the Sources page.