Frequently Asked Questions

Find answers to common questions about CarGrader scores and ratings

• What if I don't agree with the Score?
  That's fine! We've tried to account for two of the biggest variables that contribute to complaint accumulation which are time and popularity. A "Certainty" percentage has also been implemented to express our confidence in the score. However, there are many variables that are hard to model and/or predict. For example there could be contributing factors like an owner's disposition to submit a complaint might vary between luxury and economy models, or owners might have been encouraged to submit a complaint due to extensive news coverage of a vehicle recall. Both of these examples, at the moment, are hard to implement into a grading algorithm and could impact the scores here. The In some rare cases the possibility of a typo does exist, and we are resolving those typos as we find them!
• What is 'Certainty' supposed to mean?
  The Certainty percentage is meant to express our confidence in the score itself. It's primarily determined by the vehicle's reported sales figure, by calculating how easily a score is changed by a single complaint. For anyone who knows something about Statistics, sample size is essential when it comes to accuracy. Cars like a Toyota Corolla, which sell hundreds of thousands every year, have a very large sample size and the score is hardly affected by 1 or even 10 complaints and will have a high certainty percentage. But cars like an Alfa Romeo Giulia, which only sell a few thousand a year, have a small sample size and are very easily affected by a single complaint, so they get a low certainty percentage.
• What is the Score supposed represent?
  Each score is meant to represent how much a car has deviated from what the typical complaint fequency is for its age and sales volume. We have compiled 20 years of historical complaint data to determine what a normal amount of complaints is for a car that's 2, 5, or even 15 years old. The score however is not meant to reflect subjective things like performance, interior design, appearance, etc. It's primarily a measure of how problematic a car has been for owners in the past.
• Why do some cars have the same score?
  Every score in the database depends on the reported sales figures for that car. Some manufacturers will report sales for groups of models rather than specific trims or series. For example BMW sells the 330 and 340 models but will only report their sales as "3-Series". Chevrolet for example does not distinguish between the Silverado 1500, 2500, or 3500 in their sales reports - they are simply reported as "Silverado". While there are ways to work around these issues, we felt that the best way to maintain grading intergrity is to simply apply the same score to all the models in the group, because determining the distribution of sales within the group will ultimately require some guessing. And we simply didn't like the idea of guessing when our brand is founded on data-driven results.
• Why isn't there anything before 2005?
  For the majority of vehicles produced before 2005 the complaint data is often unreliable and the sales figures are usually unavailable. Not to mention, buying a car that's more than 20 years old is already taking a big chance.