I don’t know if this is just me in particular or if others have this problem as well. Let me start with I have put a lot of effort into the game into making high-end engines and running the games on that, but that’s all I can seem to do. What I mean is my ratings seem to be affected to how powerful my engine is for the game e.g. I just finished a City simulator/strategy (much like Sim City) I picked engine options that I thought would benefit the game more such as a branching story w/o advanced cutscenes, achievements, No character progression, etc. I just took out things that didn’t make any sense to have, and it seemed to affect my ratings due to those decisions. This seems to happen to me a lot, and is quite frustrating. So in my experience it seems just the more you put into the game without affecting each categories percentage quality makes it better; which of course in my opinion makes my decisions null and void. Which made me create this in hopes to discuss the topic, for maybe perhaps I’m overlooking something, or I don’t have a good enough team.
I think for each type of game, there is exactly one perfect slider-position built in the game.
The further you are from hitting that, the worse your game gets.
You can make up for bad sliders with giving the market what it wants, lots and lots of fans, new topics and great employees.
Also a perfectly slided game can fail, when you i.e. create the same game twice in a row or you let your technician design the graphics.
If the perfect slider-position doesn’t make sense to you, blame greenheart for liking different types of games than you 
Oh, and of course there is a bit of randomness in the ratings. Sometimes a mediocre title gets straight 10s and vice versa. So good luck!
Try to look at the wiki of the game, it has some useful tips.
If you want to understand how the game calculate the review score I suggest you this article