The words are semantically primitive, you're not going to define them in some other times. It might seem difficult to believe, but if you have any experience with this sort of activity it's pretty clear how it is generally easy to come to a consensus on what belongs on each category.
I sometimes deal with this with university students when they come ask me about my grades. I don't go into these details, but the truth is, I could give their papers to any other professor, and they'd get the same letter grade. It's not difficult when you're only dealing with a few categories: A, B, C, D, F. Two less than what I proposed earlier for games. An 'A' paper is quite clear to anyone in my field (philosophy). It's well written, well argued, no fallacious arguments, and it presents a novel position. A 'B' paper may not have any issues, but it could stand a number of improvements in either exposition, exploring objections, and so forth. A 'C' paper has some problems. It may contain some fallacious arguments, or may display misunderstandings. But it was a well intended attempt. A 'D' paper was probably not well intended, or if it was, it displays serious issues. An 'F' paper is just atrocious and has no redeeming qualities.
This is a common activity that happens in so many different fields, including creative ones, such as fiction writing: being able to analyze the objective factors of a work. I'm not saying the result is going to look anything like a mathematical deduction in terms of the extent and specificity of the consensus, but you can quite clearly talk in broad terms about how well or poorly written some piece of fiction is, some piece of academic scholarship is, or how well or poorly written some piece of game design is and find large consensus.
To suggest that all there is are subjective attitudes and ultimately sales numbers is incredibly pessimistic.