Basically? You can separate two concepts:
1) a good game is one which meets is objectives, or at least some set of objectives.
2) a good game is one which meets the needs of a particular person or group
The point is that there's no one set of objectives which is good. You can't say "you should design a game according to these principles" because you can't say that one game is flat out better than another. Some players will like swingy games, some will hate them, etc.
You can say that some games are good at meeting some needs, and some are better at meeting other needs. You could even theoretically prove one "better" in a limited sense if you could prove that it was better at meeting every possible need than another game.
We can definitely say "these games are good for you" in some fashion, by figuring out the things you actually are looking for in a game, and then figuring out which games deliver those things.
But you can't say "games should be designed like so", except in the most generic of higher-order criteria like "don't have rules which fight against the objectives other rules aim for". And even that game might hit the sweet spot for a limited number of people.
Basically. Or you allow the concept of "a good game but not a good game for me." I can recognize that D&D 3 does a lot of things well - they're just things I don't want. I don't think it's a bad game, but I do think it doesn't fit my needs well at all.