In order to know whether the ratings have any validity, you have to know exactly how they are compiled. I am not familiar with the CR system.
However, I have seen ratings that compare "problems per 100", that is, the number of reported problems in a given amount of time for 100 cars. What is the problem with this? It doesn't account for the type of problem.
For instance, say you have a make of car that has 15 reported problems per 100, but they are all things such as a "check engine" light which is cleared easily, squeaky brakes, or loose trim. You have another car tha has 5 problems per 100, but all of them are connecting rod through the block.
Obviously, the second car is 3 times more reliable than the first, right? However, which would you rather own?
Statistics.....without knowing the exact basis, they mean little to nothing.