The “scale” is mostly guesswork from observations done while playing the game in Legendary, and I’ve noticed all genres expect you to have at least 200 points in Gameplay in 1976, even if you’re making a Racing game, where the ‘optimal’ percentile slider for that stat is at 10%.
You’ll get more than enough points for that normally, as you’ll have game developers generating those, while graphics and sound are much harder to get earlier on, due to graphics/sound artists being stuck in the development room, where they’re still going to generate all four kinds of stats, but be horrible at three of them.
Also, a lack of experience will give a harsher review score penalty in Legendary (for the “Overall” rating) and some people say not having improvements also affects that rating, but I never use improvements and still can manage 90%+ games, so I’ll trust my own eyes more than someone else’s guesswork (and so should you, so feel free to disregard the numbers I’ll show below.)
From my own “playing with the game” runs (sandbox, unlimited money, have dozens of small teams making “the same” game and seeing what version gets what rating) I have these estimates on what numbers will get you over 90%, assuming everything else is covered (sliders, subgenre/topic, experience):
GamePlay, GRaphics, Sound, Technical
- 1976: GP: 200-250 GR: 100-150 S: 100-120 T: 100-150
- 1977: GP: 250-300 GR: 150-250 S: 130-200 T: 150-200
- 1978: GP: 350-450 GR: 220-320 S: 180-250 T: 200-350
No, I don’t have good guesses beyond that point, because these “testing” runs take hours, involve heavy micro-management (which I’m not a fan of) and a lot of RNG (which I’m less of a fan of,) and I’m not game-monogamous.
Also, to elaborate: You can get 90%+ reviews in a game even if one or two of your stats is lower than the 90% threshold, as long as it’s not too low. For example, a game with 98%-88%-81%-98% can still score 95% overall. A game with a score of 98%-70%-40%-98% will however fall under the 90% margin and will sell poorly compared to the first example. It will give you “profit” as long as you consider a couple tens of thousands of dollars profit (which I don’t.)
As a general rule: Look at the stats of your game, and the review rating it got in those stats. If one of them is below 90%, you need more of that stat. With time, you may be able to figure out a way to not even have to check. I set all my racing games’ percentile sliders to 0-50-50-0 (changing them to their ‘perfect’ 10-40-30-20 right before release,) and I get 90%+ on racing games even without a graphics or sound employee in my ranks (or even with some of my game development employees being a sickly scientist, a clueless hardware engineer and a plant-loving game tester.)