Keep in mind the classifications of these moves is also based somewhat on your rating. A lower-rated player gets credit for all sorts of things in the game review that a higher-rated player would not receive. The same thing applies for accuracy in your games. When you are lower-rated you can play all kinds of dubious moves and still have your accuracy grade really high. For instance a 1000 rated player might receive a 97% accuracy score for a game but a 2000 rated player can play the exact same moves in a 100% exactly the same game and it might give them 90% accuracy or something like that. I agree with you I don't like this ambiguity.
The same thing happens with inaccuracies, mistakes and blunders. If you have a rating around 300 you can just about hang your Queen and it might call it an inaccuracy. But if you are rated 2500 a slight imprecision may be called a blunder. They just change the criteria.
Chess By Its nature is a brutally unforgiving game. The truth always comes out in the end. I am fine with that and I would prefer if they did not try to boost my ego by grading my performance on a curve. It's an unfortunate side effect of the "everyone-gets-a-trophy" Society we live in today.
Here would be my complaint. The last couple days I have played some games with a good friend of mine who just joined and is a provisional player. Just ignore that part and concentrate on my side of the game review. He plays much better than his rating but he is considerably lower rated than myself. But that's irrelevant to my point which I'll try to make. When I look at the game review it says I have 15 best moves out of 27 for the game. There were three book moves which I consider nothing but memorization. Let's ignore those leaving 24. There are nine moves which were not the best moves. By my old school math, 15 / 24 equals .625 or 63% accuracy when generously rounding up. But the game review gives me 91%.
https://www.chess.com/analysis/game/daily/396739945
For my own purposes of self-improvement , I would much prefer they simply tell me the truth. By contrast if you look at his accuracy there were five top engine pick moves. In the real world I would call that 21% accuracy, but here on chess.com 5 / 24 = .73.
The new brilliant system implemented by Chess.com looked like it could be a major improvement to the brilliant system back then, where occasional obvious moves were given brilliant status. Well, the new system is no better, giving out brilliants for obvious sacrifices, threats, or forks. Here are three examples from my games today which made me realize how this system is broken.
Game #1:
Game 2+3:
Chess.com, find a better brilliant system.
It probably will be hard, but giving the rank of brilliant for easier to find moves is not "brilliant."